Two -dimensional superlattices of colloidal quantum dots ...lib.ugent.be/fulltxt/RUG01/002/163/541/RUG01-002163541_2014_0001_… · PbSe quantum dot synthesis PbSe semiconductor nanocrystals

Department of Inorganic and Physical Chemistry Research group Physics and Chemistry of Nanostructures

Two-dimensional superlattices of colloidal quantum dots - towards high performance

photodetectors

Thesis submitted to obtain the degree of Master of Science in Chemistry by

Willem WALRAVENS

Academic year 2013 - 2014

Promoter: prof. dr. ir. Zeger Hens

Copromoter: prof. dr. ir. Gunther Roelkens Supervisors: ir. Chen Hu and dr. Yolanda Justo

i

The author and promoter give the permission to use this thesis for consultation and to

copy parts of it for personal use. Every other use is subject to the copyright laws, more

specifically the source must be extensively specified when using from this thesis.

De auteur en promotor geven de toelating deze scriptie voor consultatie beschikbaar te

stellen en delen ervan te kopieren voor persoonlijk gebruik. Elk ander gebruik valt onder

de beperkingen van het auteursrecht, in het bijzonder met betrekking tot de verplichting

uitdrukkelijk de bron te vermelden bij het aanhalen van resultaten uit deze scriptie.

Ghent, June 2014

The promotor The author

Prof. dr. ir. Zeger Hens Willem Walravens

Preface

Dear reader,

this thesis is the culmination of a year dedicated to experimental laboratory work. As a

last part of the master’s degree, this represents the transition from the classroom, where

the student is provided with clear-cut answers, to the laboratory, where questions and

possibilities reign. This rather different perspective is very exciting indeed, but it is often

a process of trial and error. Therefore I would like to thank my promoter, Zeger Hens, and

both my supervisors, Chen Hu and Yolanda Justo, for their guidance throughout the year

and for sharing their experience and insights. I would also like to thank Katrien Haustraete

for taking TEM images and Stijn Flamee for taking SEM images and introducing me to

the technique. Furthermore, all people who have helped me in one way or another this

year have my gratitude. I hope you enjoy this thesis as much as I did performing the

experiments and writing down the story of the results.

Willem Walravens

Ghent, June 2014

v

Dutch summary

Deze thesis onderzoekt de mogelijkheid om PbSe quantum dot (QD) superstructuren te

gebruiken als actief sensormateriaal voor fotodetectie in het nabij-infrarood gedeelte van

het elektromagnetisch spectrum. De QDs werden gesynthetiseerd via de zogenaamde hot

injection methode. Vervolgens werden elektronisch gekoppelde QD structuren gevormd

via dropcasting op een vloeibare subfase en via Langmuir-Schaefer depositie. Verschillende

morfologieen werden verkregen door verandering van de samenstelling en temperatuur van

de subfase, en het tempo waarmee de QDs werden toegevoegd. De structuren werden

gekarateriseerd via TEM, SEM, AFM en FTIR. De fotogeleidende eigenschappen werden

onderzocht via I-V metingen van de lagen, afgezet op gouden elektroden. Alle gekoppelde

superstructuren vertoonden fotogeleiding en de resultaten wijzen verder op het belang van

de QD oppervlakte chemie en gecontroleerde trap state introductie voor het verbeteren

van de prestatie van de fotodetectoren.

vii

Quantum Dot Superstructures for Photodetection in the Near-Infrared

W. Walravensa,b, C. Hua,b, Y. Justoa, G. Roelkensb, Z. Hensa

a Physics and Chemistry of Nanostructures, Department of Inorganic and Physical Chemistry, Ghent University, 9000 Ghent, Belgium

b Photonics Research Group, INTEC Department, Ghent University, 9000 Ghent, Belgium

This paper reports the use of PbSe quantum dot (QD) superstructures as active sensing material for photodetection in the near-infrared. Electronically coupled QD superstructures were obtained via dropcasting on a liquid subphase and via Langmuir-Schaefer deposition. Different morphologies were obtained by changing the composition and temperature of the subphase, and the rate at which the QDs were added. The layers were characterized by TEM, SEM, AFM and FTIR. The photoconductive properties were determined via I-V measurements of the layers deposited on gold comb-like electrodes. All coupled superstructures show a photocurrent and the results point to the importance of the QD surface chemistry and directional trap state engineering as factors for enhancing the device performance.

Introduction

Colloidal quantum dots (QDs) are semiconductor nanocrystals which are dispersed as colloids in solution. Due to their dimensions below the exciton Bohr radius they exhibit a size dependent electronic structure via the quantum confinement effect. Quantum confinement translates in increasing band gap energies with decreasing crystal size, and thus enables spectrally tunable absorption spectra via careful control of the size. Their relatively cheap and easy synthesis makes them further very useful as building blocks for novel optoelectronic devices.

Colloidal QDs are synthesized via a bottom-up approach in a controlled crystallization of the semiconducting material. This is typically carried out in an organic solvent where the nanocrystals are stabilized with suitable organic ligands bound to the inorganic core. Colloidal dispersions make it possible to process QDs via solvent-based deposition techniques such as dropcasting, dipcoating, Langmuir-Blodgettry etc. These techniques have the advantage of being cheap, easy and fast compared to conventional vacuum- based techniques, which require complex and expensive equipment. However, production of high quality films via solvent-based techniques is an ongoing research and does not yet reach the standards of films produced by vacuum-based techniques.

Since QDs are light absorbing materials, they can be used as the active sensing material in photodetectors. More specifically, a photoconductor type of detector is more appropriate than a photodiode type of detector, as it makes better use of some of the advantages QDs offer. The operating principle of a photoconductor is schematically shown in Figure 1. A semiconducting sensing layer is placed between two biased electrodes. When a photon is

ix

absorbed, an electron is promoted to the conduction band, increasing the conductivity of the semiconducting layer. The rise in conductivity is detected as a photocurrent in the external circuit.

A first advantage of using QDs as sensing material is that they have a much larger surface area than bulk material. Since surface defects or chemical functionalization of the surface introduces trap states in the mid-gap region, electrons are trapped for a longer time and subsequently the holes can drift for a longer amount of time. In the case that a hole travels through the external circuit multiple times while the electron is trapped, a condition of photoconductive gain is created. This means that much larger photocurrents are generated per absorbed photon and thus the control over trap states provides a route to produce highly sensitive photodetectors.

A second advantage is that QDs exhibit multiple exciton generation (MEG) at lower photon energies than the bulk material (relative to the band gap energy Eg) [1]. MEG is a process in which a high energy photon is absorbed, at least twice the band gap energy. The excited electron can relax to the edge of the conduction band by releasing a photon with energy Eg. This photon is absorbed by a second electron, thereby creating a second exciton. Although this is an important advantage of QDs, it is less important for QD photodetectors in the infrared since it involves photon energies outside the region of interest.

Figure 1. Operating principle of a photoconductor. After absorption of a photon, an electron is excited to the conduction band (CB). The created hole in the valence band (VB) generates a photocurrent by circulating through the external circuit. Trapping of the electron in a trap state (TS) increases the lifetime of the exciton and thereby increases the photocurrent. After relaxation to the ground state, the photocurrent falls back to 0.

A functioning photoconductor however, requires charge transport between the electrodes. Since as-synthesized QDs have a relatively large insulating ligand shell, they do not allow charge carriers to be transported between the separate QDs. A procedure is thus required to remove the native ligand shell and induce an electronic coupling between QDs, thereby enabling a high mobility for photo-induced charge carriers. Several post-treatment strategies are found in literature, where interdot connections are formed after an initial layer formation. Such strategies include annealing [2, 3], ligand exchange [4-8] or ligand displacement [9, 10]. However, these strategies have the disadvantage of creating large cracks in the layer, which is detrimental for their conductivity. Pre-treatment includes ligand exchange prior to layer formation [11, 12]. This is a good strategy since it avoids crack formation, but a stable dispersion of inorganically capped PbSe QDs has not been reported yet. Other strategies influence the QD ligand shell at the moment the superlattice is formed. This is conveniently done by forming the superlattice on an immiscible, liquid subphase. On one hand this enables the addition of extra reagents and

x

on the other hand it acts as a reservoir for the removed ligands. This method has successfully yielded connected QD superlattices by either adding formic acid to an acetonitrile subphase, or by forming the superlattice on ethylene glycol at elevated temperature [13, 14].

This paper intends to show the usability of connected QD superlattices as a photosensitive layer in a photoconductor type of device. To this end, superlattices were formed on a liquid subphase either by dropcasting or by using the Langmuir-Schaefer technique. Different morphologies were obtained by changing the composition and temperature of the subphase. Their photodetection capabilities were investigated via I-V measurements, and indicate that both the morphology and surface chemistry play an important role in the device performance.

Experimental PbSe quantum dot synthesis

PbSe semiconductor nanocrystals were synthesized by a method found in literature [14]. The crystals were grown at 180 °C for 60 s. After precipitation and purification of the obtained product, the QDs were dispersed in either toluene or hexane.

PbSe superlattice formation

Dropcasting on ethylene glycol. In this method, 10-50 µl of a PbSe QD solution with concentration varying from 0.29 µM to 2.9 µM, was cast onto 1 ml ethylene glycol, in a glass container with 10 mm diameter. Samples of the film were taken at the center of the vial after 1 h. After deposition, the substrate was immersed in deionized water to remove residual ethylene glycol and dried under a nitrogen flow. All the dropcasting experiments were carried out in a nitrogen purged glovebox.

Langmuir-Schaefer deposition. In this technique, 100 µl of a filtered 2.46 µM QD solution in hexane was added drop-wise to the Langmuir trough. After the solvent evaporated, the QDs were compressed to a desired pressure, by reducing the available surface area with 10 cm²/min.

Device fabrication

Photodetectors were prepared by patterning gold comb-like electrodes on a Si/SiO2 substrate via photolithography. The photosensitive layers were placed on top of the electrodes by direct stamping of the superlattices formed on the subphase. I-V measurements

I-V curves were obtained by sweeping a voltage from -5 V to +5 V across the electrodes and recording the current. The current was measured under different illumination levels, ranging from 0 mW to 800 mW, with a laser operating at a fixed wavelength of 1550 nm. From the resulting photocurrents, i.e. measured current minus dark current, responsivities (A/W, calculated as R=Iphoto/Pincident) and contrast ratios (Iphoto/Idark) were plotted.

xi

Results

PbSe QD synthesis The TEM image in Figure 2(a) shows that the synthesis yielded monodisperse, quasi-spherical PbSe nanocrystals. The size of the crystals was determined at 6.0 nm ± 4.25 % from the first exciton peak in the absorbance spectrum (Figure 2(b)).

Figure 2. (a) TEM image of the as-synthesized PbSe QDs. (b) Absorption spectrum corresponding to the QDs in (a). PbSe QD superlattices

Superlattices by dropcasting on ethylene glycol Addition of Na2S to the subphase. Figure 3(a-d) shows the microstructure of

monolayers formed by dropcasting 18 µl of a 0.29 µM solution at room temperature. From left to right, the amount of Na2S in the subphase increases from naught to a 100 fold excess, compared to the total amount of ligands in the QD solution.

The addition of Na2S to the subphase increased the amount of interparticle connections and eventually resulted in a dense, connected superlattice of QDs. The FTIR spectra in Figure 3(e) show a strong reduction of C-H stretch vibrations by increasing the amount of Na2S in the subphase. This indicates that Na2S successfully removed the native oleate ligands and that the QD surface becomes passivated with S2- ions.

The long range morphology of the obtained superlattices was studied with SEM (Figure 3(f-g)). The images show that the superlattices formed by the addition of a 10 fold (A) and a 50 fold (B) excess Na2S are very homogeneous with almost no cracks over an area of several tens of square micrometer.

xii

Figure 3. (a-d) TEM images of the superlattices formed by dropcasting on ethylene glycol. The amount of Na2S added to the subphase is no Na2S (a), a 10 fold excess (b), a 50 fold excess (c) and a 100 fold excess (d). Scale bars are 10 nm. (f-g) SEM images of superlattices formed with a 10 fold (f) and a 50 fold excess (g) Na2S. Scale bars are 1 µm. (e) FTIR spectra showing the decrease of C-H stretch vibrations by the addition of Na2S.

Temperature of the subphase. The effect of temperature on the formation of QD superlattices was studied by heating the subphase to 50 °C before adding the QD solution. Figure 4(a-b) shows the effect of adding 50 µl of a 1.4 µM QD solution at room temperature (a) and at 50 °C (b). At room temperature, the nanocrystals arranged in a hexagonal close packed structure and the distance between the crystals indicate that the ligand shell was not altered in a significant way. On the other hand, at elevated temperature, the individual nanocrystals merged together and formed a superlattice of connected QDs with a quasi-cubic symmetry. The FTIR spectra in Figure 4(d) confirm that the majority of oleate ligands are removed.

This procedure yielded a thicker, less homogeneous film. The surface profile in Figure 4(e) follows the line drawn in the AFM topography image in Figure 4(f), and shows that the superlattice mainly consists of a bilayered structure alternated with smaller monolayer regions. Figure 4(c) is an SEM image showing that the layer contains a number of small cracks with widths in the µm range.

xiii

Figure 4. (a-b) TEM images of the superlattices formed by dropcasting a QD solution on ethylene glycol at room temperature (a) and at 50 °C (b). Scale bars are 50 nm. (c) SEM image showing crack formation at 50 °C. Scale bar is 1 µm. (d) FTIR spectra showing the decrease of C-H stretch vibrations in the superlattice shown in (b). (e) Surface profile following the line drawn in the topography image (f).

Rate of QD addition. The previously shown superlattices were obtained by adding the QD solution in one swift injection. A completely different morphology was obtained by adding the QDs in a drop-wise fashion to the subphase at 50 °C (Figure 5(a)). Instead of an assembly of individually connected QDs, the slow addition yields a structure where the QDs fuse together and form a molten-like network. Similar to the previous results, the native oleate ligands are successfully removed, as indicated by the removal of C-H stretch vibrations in the FTIR spectrum (Figure 5(b)). The AFM topography image shows that the layer has a thickness around 11 nm (Figure 5(c-d)).

Figure 5. (a) TEM image of the structure formed by drop-wise addition of the QDs at 50 °C. (b) FTIR spectra showing the decrease of C-H stretch vibrations in the superlattice shown in (a). (c) Surface profile following the line drawn in the topography image (d).

Superlattices by Langmuir-Schaefer deposition. In a first series of experiments with this technique, it was established that using water or ethylene glycol as subphase resulted in hexagonal close packed layers where the QDs retained their native ligand shell. On the other hand, by using diethylene glycol as subphase, the QDs instantaneously formed interparticle connections. Since the irreversible bonds reduced the mobility of the

xiv

nanocrystals, it was impossible to compress the layer to yield a homogeneous, dense structure. To avoid this, two modifications were made to the subphase.

A first modification was using a subphase consisting of 30% diethylene glycol and 70% ethylene glycol. This drastically improved the quality of the layer, and homogeneous areas of several square centimeters were obtained. Figure 6(a) show that although the density is not maximal, the spread of the nanocrystals is relatively homogeneous. Figure 6(b) more clearly shows that connections are formed between QDs, although a significant amount is still separated by an oleate ligand shell.

The second modification was adding a 100 fold excess Na2S to ethylene glycol. A relatively dense layer was obtained where the nanocrystals have a rather random, non-symmetrical stacking (Figure 6(c)). A large fraction of the QDs show interparticle connections (Figure 6(d)), but also many nanocrystals are not connected, signaling an incomplete ligand removal. Although Na2S is a good reagent for oleate removal, the large volume and surface area of the subphase means that less S2- ions are available at the subphase-QD interface and thus less oleate ligands are exchanged.

The FTIR spectra in Figure 6(g) confirm that both modifications result in only a partial removal of the oleate ligands.

As for the long range morphology, SEM images indicate that reasonably homogeneous films were obtained (Figure 6(e-f)), considering the difficulty to compress layers where QDs form interparticle connections.

Figure 6. (a-b) TEM images of the structure formed by Langmuir-Schaefer deposition on an EG subphase containing 30% DEG. Scale bars are 100 nm (a) and 10 nm (b). (c-d) TEM images of the structure formed by Langmuir-Schaefer deposition on an EG subphase containing a 100 fold excess Na2S. Scale bars are 10 nm (c) and 5 nm (d). (e) SEM image corresponding to the layer shown in (a) and (b). (f) SEM image corresponding to the layer shown in (c) and (d). (g) FTIR spectra of the Langmuir-Schaefer films showing a partial reduction of the C-H stretch vibrations.

xv

PbSe QD photodetectors

Addition of Na2S to the subphase. The increase of interparticle connections directly enhanced the conducting behavior of the superlattice. This is illustrated by an increase of the dark current (at 5 V) from 0 µA to 10.5 µA going from no Na2S to a 50 fold excess, and an increase from 10.5 µA to 29.0 µA going from a 50 fold to a 100 fold excess. The increased conductivity allows the superlattice to produce more photocurrent per incident photon power. This resulted in a significant increase of both photocurrent and responsivity by the addition of more Na2S (Figure 7(a-b)). The comparable contrast ratio (Figure 7(c)) suggests that the observed increase in photocurrent and responsivity is mainly due to the increased conductivity and not due to an additional increase of the carrier lifetime.

Figure 7. Photocurrent (a), responsivity (b) and contrast ratio (c) at 5 V in function of illumination power, as obtained from the measured I-V curves. The curves correspond to the structures shown in Figure 3(c) (red curve) and Figure 3(d) (blue curve).

Temperature of the subphase. Fast addition of QDs. This structure (see Figure 4(b)) shows a remarkable increase of

conductivity, with a dark current (at 5 V) of 209.8 µA. On one hand, this is due to a structure with a high density of interconnected nanocrystals. On the other hand, the bilayered structure possibly serves as a bridge between cracks or isolated parts, thereby reducing the percolation path and activating 'dead' areas, which are otherwise effectively lost in a monolayer. The increased conductivity led to a significant increase in photocurrent and correspondingly, relatively high responsivities were obtained (Figure 8(a-b)). The contrast ratio is shown in Figure 8(c).

Figure 8. Photocurrent (a), responsivity (b) and contrast ratio (c) at 5 V in function of illumination power, as obtained from the measured I-V curves. The curves correspond to the structure shown in Figure 4(b).

xvi

Drop-wise addition of QDs. The molten-like structure obtained by drop-wise addition of the QDs shows a dark current (at 5V) of 35.9 µA. Although the electrical connections in this structure are good, both the lower density and the more hindered percolation path contribute to a lower dark current. The photocurrents, responsivities and contrast ratios are shown in Figure 9(a-c).

Figure 9. Photocurrent (a), responsivity (b) and contrast ratio (c) at 5 V in function of illumination power, as obtained from the measured I-V curves. The curves correspond to the structure shown in Figure 5(a).

Langmuir-Schaefer films. The photodetectors produced by 4 Langmuir-Schaefer depositions have a very low dark conductivity, with dark currents (at 5 V) of 0.69 µA on the EG/DEG subphase and 0.79 µA on the Na2S containing subphase. Upon illumination, photocurrents of a few microamperes are generated, resulting in relatively low responsivities (Figure 10(a-b)). However, due to the low dark current, high contrast ratios were obtained in these structures (Figure 10(c)).

Figure 10. Photocurrent (a), responsivity (b) and contrast ratio (c) at 5 V in function of illumination power, as obtained from the measured I-V curves. The curves correspond to the structures shown in Figure 6(a) (red curve) and Figure 6(c) (blue curve).

Discussion The role of S2-

By taking into account the density and thickness, a responsivity per QD of 6 10-11 A/W is obtained for the S2- passivated structure (Figure 3(d)) while the non-S2- passivated structure (Figure 4(b)) yields 13 10-11 A/W per QD. It is seen that this quantity is about twice as high. At first sight this can be explained by the better conductivity of the latter, but since the dark current is about 7 times higher, another factor plays in favor of the first. If the same level of doping and quantum efficiency is assumed in the QDs, the responsivity scales with τlifetime/τtransit, meaning that the lifetime of the carrier is about 3.5

xvii

times higher in the S2- passivated structure. This indicates that the S2-ions possibly introduce effective electron traps in the band structure. The role of surface area and morphology

Since the structures in Figure 4(b) and 5(a) are obtained by a similar procedure, it is assumed that they have a similar surface chemistry and that trap states are mainly introduced by surface defects. In this perspective, the higher contrast ratio in the molten structure can be rationalized by both a larger surface area, with possibly more surface defects, and a more complete removal of the organic ligands (see FTIR spectra in Figure 4(d) and 5(b)). Additionally, the complex morphology of the molten-like structure can possibly further increase the carrier lifetime by a principle of harder-to-find recombination centers. The role of surface passivation

Both Langmuir-Schaefer films show a very low dark current. However, upon illumination, photocurrents with relatively high contrast ratios are generated, meaning that a conductive path is available for the charge carriers. This indicates that few intrinsic mobile charge carriers are present in the structure. Since both Langmuir-Schaefer films still contain a large amount of native ligands, it is well possible that the oleate provides a better passivation and that in the other methods the many created surface states lead to an effective doping of the QDs, thereby increasing the dark current. This suggests that a superlattice with a well passivated surface, followed by controlled introduction of non-doping trap states could further increase the device performance.

Conclusion

In conclusion, this paper showed that several strategies successfully removed the native oleate ligand shell. This enabled the fabrication of nanometer sized, interconnected QD superlattices over a large area. The reported procedures include the addition of Na2S or diethylene glycol to the subphase, and increasing the temperature of the subphase.

The obtained structures furthermore showed promising features towards their implementation as sensitive layers in photoconductors. The results point to several important factors influencing the device performance, such as surface passivation, the nature of the trap states, carrier mobility, and morphology.

Acknowledgements

The research was supported by the Department of Inorganic and Physical Chemistry

and the INTEC Department of Ghent University. The authors would like to express their gratitude to Katrien Haustraete for taking TEM images and Stijn Flamee for taking SEM images.

References

1. Nozik, A.J., Nanophotonics: Making the most of photons. Nat Nanotechnol, 2009.

4(9): p. 548-9.

xviii

2. Wills, A.W., et al., Thermally Degradable Ligands for Nanocrystals. ACS Nano, 2010. 4(8): p. 4523-4530.

3. Simon, P., et al., Interconnection of Nanoparticles within 2D Superlattices of PbS/Oleic Acid Thin Films. Advanced Materials, 2014: p. n/a-n/a.

4. Talapin, D.V. and C.B. Murray, PbSe Nanocrystal Solids for n- and p-Channel Thin Film Field-Effect Transistors. Science, 2005. 310(5745): p. 86-89.

5. Hu, C., et al. Short-wave infrared colloidal quantum dot photodetectors on silicon. 2013.

6. Jarosz, M.V., et al., Photoconductivity studies of treated CdSe quantum dot films exhibiting increased exciton ionization efficiency. Physical Review B, 2004. 70(19): p. 195327.

7. Oh, S.J., et al., Designing High-Performance PbS and PbSe Nanocrystal Electronic Devices through Stepwise, Post-Synthesis, Colloidal Atomic Layer Deposition. Nano Letters, 2014. 14(3): p. 1559-1566.

8. Rosen, E.L., et al., Exceptionally mild reactive stripping of native ligands from nanocrystal surfaces by using Meerwein's salt. Angew Chem Int Ed Engl, 2012. 51(3): p. 684-9.

9. J. Am. Chem. Soc. 2012, 20705−20712Baumgardner, William J., K. Whitham,

and T. Hanrath, Confined-but-Connected Quantum Solids via Controlled Ligand Displacement. Nano Letters, 2013. 13(7): p. 3225-3231.

10. Hanrath, T., et al., PbSe Nanocrystal Network Formation during Pyridine Ligand Displacement. ACS Applied Materials & Interfaces, 2009. 1(2): p. 244-250.

11. Nag, A., et al., Metal-free inorganic ligands for colloidal nanocrystals: S2-, HS-, Se2-, HSe-, Te2-, HTe-, TeS3(2-), OH-, and NH2- as surface ligands. J Am Chem Soc, 2011. 133(27): p. 10612-20.

12. Dong, A., et al., A Generalized Ligand-Exchange Strategy Enabling Sequential Surface Functionalization of Colloidal Nanocrystals. Journal of the American Chemical Society, 2010. 133(4): p. 998-1006.

13. Dong, A., Y. Jiao, and D.J. Milliron, Electronically Coupled Nanocrystal Superlattice Films by in Situ Ligand Exchange at the Liquid–Air Interface. ACS Nano, 2013. 7(12): p. 10978-10984.

14. Evers, W.H., et al., Low-Dimensional Semiconductor Superlattices Formed by Geometric Control over Nanocrystal Attachment. Nano Letters, 2013. 13(6): p. 2317-2323.

xix

Contents

Preface v

Dutch Summary vii

Article ix

1 Introduction 1

1.1 Colloidal Quantum Dots . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

1.1.1 Quantum confinement . . . . . . . . . . . . . . . . . . . . . . . . . 2

1.1.2 Light absorption by quantum dots . . . . . . . . . . . . . . . . . . . 4

1.1.3 Colloidal synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

1.2 Superlattices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

1.2.1 Thermodynamic considerations . . . . . . . . . . . . . . . . . . . . 6

1.2.2 Superlattices in practice . . . . . . . . . . . . . . . . . . . . . . . . 8

1.3 Application: photodetectors . . . . . . . . . . . . . . . . . . . . . . . . . . 11

1.3.1 Photoconductors . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

1.3.2 Photodiodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

1.4 Photoconductors based on quantum dot superlattices . . . . . . . . . . . . 14

2 Quantum Dot Synthesis 15

2.1 Synthesis of PbSe QDs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

2.2 Results and discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

xxi

Contents xxii

3 Quantum Dot Superlattices 20

3.1 Superlattices by dropcasting on ethylene glycol . . . . . . . . . . . . . . . 20

3.1.1 Addition of Na2S to the subphase . . . . . . . . . . . . . . . . . . . 21

3.1.2 Temperature of the subphase . . . . . . . . . . . . . . . . . . . . . 25

3.1.3 Rate of QD addition . . . . . . . . . . . . . . . . . . . . . . . . . . 29

3.2 Superlattices by Langmuir-Schaefer deposition . . . . . . . . . . . . . . . . 34

3.3 Light absorption by superlattices . . . . . . . . . . . . . . . . . . . . . . . 40

3.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

3.4.1 Oleate and the subphase . . . . . . . . . . . . . . . . . . . . . . . . 42

3.4.2 Oriented attachment . . . . . . . . . . . . . . . . . . . . . . . . . . 43

3.4.3 Langmuir films . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

3.4.4 Absorption enhancement . . . . . . . . . . . . . . . . . . . . . . . . 46

3.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

4 Quantum Dot Photodetectors 49

4.1 Addition of Na2S to the subphase . . . . . . . . . . . . . . . . . . . . . . . 50

4.2 Temperature of the subphase . . . . . . . . . . . . . . . . . . . . . . . . . 51

4.3 Superlattices by Langmuir-Schaefer deposition . . . . . . . . . . . . . . . . 55

4.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

4.4.1 Responsivity per QD . . . . . . . . . . . . . . . . . . . . . . . . . . 57

4.4.2 Fast versus slow addition . . . . . . . . . . . . . . . . . . . . . . . . 57

4.4.3 Langmuir-Schaefer films . . . . . . . . . . . . . . . . . . . . . . . . 58

4.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

A QD Langmuir-Blodgett/Schaefer films 60

Bibliography 62

Chapter 1

Introduction

The research towards implementation of nanotechnology into integrated circuitry is a very

interesting and promising topic. The emerging field of photonics for example, aims at using

the peculiar interactions of nanocrystals with light to build devices which process photon

fluxes, much like today’s electronics process electrical signals. An important advantage is

that signals travel at the speed of light, making the devices possibly much faster than their

electronic counterpart. The combination with nanotechnology, resulting in faster, smaller

and greener devices, seems like a good fit for future technology. Crystals with sizes in the

nanometer range are easily prepared and scientists can use them as building blocks in a

multitude of devices. However, the fabrication of accurate, reliable and reproducible devices

in this direction demands a lot of research, not at least since matter behaves differently at

the nanoscale and formerly unknown effects can come into play.

This introduction first discusses nanocrystals of semiconducting material, i.e. quantum

dots (QDs), and how they behave differently from their bulk form, both electronically and

optically. In a second part, the organization of QDs in a structured manner, a so called

superlattice, is discussed. A thermodynamic point of view is given, and a review of the

state-of-the-art in superlattice formation is provided. The third part discusses photode-

tectors in general. Different types and their working principles are briefly summarized.

Finally, these three pillars are used to discuss the subject of this thesis; the fabrication of

1

Chapter 1. Introduction 2

a photodetector based on a QD superlattice.

1.1 Colloidal Quantum Dots

1.1.1 Quantum confinement

The electronic structure of a semiconducting material typically consists of two energy

bands: a filled or quasi-filled valence band, and an empty or quasi-empty conduction band.

Between these bands is a zone without energy levels, with an energy difference called the

band gap energy Eg. An exciton, i.e. a bound electron-hole pair, is created if an energy

equal to or larger than Eg is absorbed by an electron. This is the most applied property

of semiconductors.

The problem of an electron traveling a crystal can be studied by quantum mechanics as a

free particle in a box. In a first approximation, the boundaries of the box at x = 0 and

x = a can be seen as imposing an infinite potential barrier. However, for mathematical

convenience, the ensuing condition that the wavefunction should vanish at these boundaries

is often replaced by a periodic boundary condition. Here, it is required that the values

at x = 0 and x = a are identical. In 3 dimensions, the eigenfunctions and eigenenergies

obtained by solving the Schrodinger equation are represented by:

ψ(r) =

√8

a3/2sin(

nxπ

ax)sin(

nyπ

ay)sin(

nzπ

az) (1.1)

E =~2π2

2ma2(n2

x + n2y + n2

z) (1.2)

Important in this result is that the eigenenergies are proportional to 1/a2, meaning that

the eigenenergy (i.e., the kinetic energy) of an electron strongly increases with decreasing

size of the crystal. This increase can be seen as an extra energy cost in order to confine

an electron to a smaller space. In the end, the quantum confinement effect translates as

the presence of discrete energy levels near the edges of the valence and conduction band,

with the exact position of the energy levels, and thus the value of the band gap, being

determined by the size of the crystal.


In this context, a quantum dot is defined as a solid material where this confinement takes

place in three dimensions, and as a consequence, a QD behaves as something between an

atomic species and a bulk material. Their intermediate electronic structure is illustrated

in Figure 1.1.

One way to make the confinement effect observable in a semiconducting material is by

reducing the crystallite size to below the exciton Bohr radius. This is the most probable

distance between the hole and electron in an exciton and is given by:

aeh = εm

µeh

ab (1.3)

with ab the Bohr radius, m the mass, µeh the reduced mass of the exciton and ε the dielectric

constant. Depending on the material, the exciton Bohr radius in semiconducting materials

can be relatively large (up to 100 nm) due to a small reduced exciton mass or a large

dielectric constant. With the correct synthetic bottom-up approach, these crystallite sizes

can be easily obtained, and this has the advantage that quantum confinement is accessible

at room temperature.

Since the energy of an exciton is determined by the band gap energy, which in turn is

determined by the size of the crystal, quantum confinement can be exploited to tune the

absorption and emission spectra near the band gap to a desired value.

Figure 1.1: A schematic representation showing the electronic structure of a bulk semiconduc-

tor (left), a quantum dot (middle), and a single molecule (right). The electronic

structure of a QD is in between that of a bulk material and a single molecule.


1.1.2 Light absorption by quantum dots

When a photon flux is incident on a thin, homogeneous film of bulk semiconducting mate-

rial, the light beam is attenuated by absorption of photons. Equation 1.4 gives the intensity

of the beam after it has traveled a distance x through the film. In this equation, κ repre-

sents the extinction coefficient and λ the wavelength of the incident light. The attenuation

follows an exponential decay, and the characteristic decay length in a log10 scale is defined

as the absorption coefficient α of the bulk material (Equation 1.5).

I(x) = I0e− 4πκ

λx (1.4)

α =ln10A

L=

4πκ

λ(1.5)

For a colloidal solution of QDs, the intrinsic absorption coefficient of the QDs µi is related

to α by Equation 1.6 [15], where µi is obtained by dividing the absorption coefficient of the

composite (µ) with the volume fraction (f ) of QDs. On one hand, the intrinsic absorption

of QDs is enhanced by a factor n/ns, where n and ns are respectively the refractive indices

of the bulk material and the solvent. On the other hand, the intrinsic absorption is reduced

by the square of the local field factor fLF (0<fLF<1). This local field factor arises from a

partial screening of the electrical field of the incident photons, and can be rationalized as

follows. Since their diameter is much smaller than the wavelength of light, QDs subjected

to a photon flux can be described as small spheres in a homogeneous electrical field (E0).

This induces an opposing internal electrical field (Eint) in the QD, the size of which depends

on the depolarization factor of the material. This internal field effectively reduces the local

electrical field (Eloc) sensed by the QD, resulting in a smaller absorption coefficient. This

is schematically shown in Figure 1.2.

µi =µ

f=

n

ns

|fLF |2α (1.6)


Figure 1.2: A schematic representation showing that the electrical field sensed by the QD (Eloc)

is smaller than the incident field (E0) by induction of an internal opposing electrical

field in the QD (Eint).

1.1.3 Colloidal synthesis

Colloidal QDs can be synthesized at lab scale through a bottom-up approach. A well-

known method is the hot injection synthesis, where an appropriate precursor is heated to

its decomposition temperature, at which moment a solution containing the counter ion is

rapidly injected. The aim is to create very quickly a highly supersaturated mixture that

causes a sudden burst of nucleation. The nuclei start to grow and both nucleation and

growth relax the supersaturated situation. In a first instance, growth occurs via diffusion

of the ions in solution. However, as the concentration of ions decreases, growth by Ostwald

ripening becomes more important. This is a process in which large crystals grow larger at

the expense of ions and small crystals, which dissolve and redeposit onto larger crystals.The

timing of the growth process thus determines the average size and size distribution of the

crystals.

Since this reaction effectively leads to the formation of solid material, the formed crystals

have a tendency to precipitate. In order to prevent this, a surfactant with a favorable

solvent interaction is added to the reaction medium. This can be for example an amine

or carboxylic acid with a relatively long carbon chain. They are usually introduced as the

counter ion in the metal precursor solution. After quenching of the reaction and purification

of the reaction mixture, a colloidal solution of QDs, consisting of an inorganic core with


an organic capping shell is obtained.

1.2 Superlattices

1.2.1 Thermodynamic considerations

From a thermodynamic point of view, the formation of a superlattice can take place at con-

stant temperature and pressure when the total Gibbs free energy of the system decreases.

The driving force can be energetic, entropic or a combination of both, depending on the

type of interactions that play a role. The change of Gibbs free energy can be written in

terms of enthalpy and entropy as

∆Gsys(t) = ∆Hsys(t)− T∆Ssys(t) (1.7)

where t denotes the evolution of the superlattice formation. The Gibbs free energy can be

minimized by minimizing the enthalpy and/or by maximizing the entropy. Possible interac-

tions contributing to the enthalpy can be charge-charge, charge-dipole, dipole-dipole, and

Van der Waals interactions. The entropy is related to the number of possible microstates

within an ensemble of particles. Different systems can be distinguished, depending on the

thermodynamic driving forces that occur in the superlattice formation.

In a first system, the formation of a superlattice can be seen as a crystallization of non-

interacting hard spheres. This means that the contribution of the enthalpy is zero, and

the crystallization is entropy driven. At first sight, this seems counter intuitive, since crys-

tallization from solution introduces order to the system. However, it has been found that

in concentrated solutions (volume fractions above 49%), the free volume per nanocrystal

in an ordered lattice is larger than the free volume per nanocrystal in the disordered liq-

uid, and this compensates the loss in conformational freedom [8,9]. For a stable colloidal

dispersion, the hard sphere model can be applied for two reasons: the Van de Waals interac-

tions between the nanocrystals are effectively screened by the solvent, and the short-range

attraction between the inorganic cores is too weak to lead to aggregation (due to steric

hindrance of the ligands).


For systems where the enthalpy contributes to the free energy, different scenarios can take

place. For example, larger and anisotropic nanocrystals have stronger, directional Van der

Waals interactions and in binary, spherical nanocrystals with an inhomogeneous distri-

bution of elements, dipole moments can arise. When these forces extend over the ligand

molecules, the free energy during crystallization can be lowered by arranging certain facets

along the direction of these forces. This leads to a preferential orientation of nanocrystals

in the superlattice. These forces are electrostatic, so the nature of the solvent has a big in-

fluence, since a high dielectric constant more effectively screens Coulomb and dipole-dipole

interactions [30,29].

The formation of a superlattice can also be purely enthalpy driven. In this case chemical

bonds are formed between the nanocrystal building blocks and when the atomic bonding

happens via specific crystal facets, a process called oriented attachment takes place. The

big difference with self-assembled layers is the strength of the interactions: irreversible,

chemical bonds versus entropic and Van der Waals interactions. An important requirement

for oriented attachment is that the facets involved in attachment are available for bonding

and are not covered with bulky ligands, as is usually the case after synthesis.

A more practical way to look at superlattice formation is by considering the condition that

the chemical potentials (µ) of the different phases should be equal at equilibrium. Since the

chemical potential of nanocrystals in a stable dispersion is lower than that of nanocrystals

in a superlattice, a phase transition can be induced by increasing the chemical potential

of the colloidal dispersion. Equation 1.8 expresses the chemical potential in terms of the

standard chemical potential and concentration, and provides two possible ways to increase

µ. The first is to increase the concentration, for example by evaporation of the solvent.

The second is to increase the standard chemical potential. This can be done by addition of

a miscible non-solvent, which effectively increases the potential energy of the nanocrystals

in solution.

µ = µo + kBT lnc

c0(1.8)


1.2.2 Superlattices in practice

When using as-synthesized colloidal nanocrystals, i.e. nanocrystals with an inorganic core

and an organic shell, the easiest way to form a superlattice is by letting the crystals

self-assemble during evaporation of the solvent. The quasi-spherical nanocrystals arrange

themselves in a hexagonal close packed structure, which has the highest packing density.

Evaporation directly on a substrate has been extensively used to form self-assembled su-

perlattices consisting of metallic or semiconducting nanocrystals [22,3]. It is also possible to

form binary superlattices by introducing different sizes of nanocrystals, or by combining

metallic with semiconducting nanocrystals [26,19,28]. Two examples of self-assembled super-

lattices are shown in Figure 1.3. The figure illustrates that binary superlattices can form

many more conformations than the typical hexagonal packing, resembling more atomic

and ionic lattices.

Figure 1.3: (A) A self-assembled superlattice of 5 nm Au nanocrystals showing a hexagonal

close packed structure. (B) A binary superlattice with a AB13 unit cell, where A

are 11 nm γ − Fe2O3 nanocrystals and B 6 nm PbSe nanocrystals.

Another method to form monolayers, i.e. 2D superlattices, is by evaporating a dilute

solution on an immiscible liquid substrate and gently pushing the layer together. This is

known as the Langmuir technique and can yield very homogeneous monolayers over a large

area [18,20,32,1] (see Appendix A).

Although a wide variety of superlattices can be made with these techniques, a major draw-

back is that the layers are electrically insulating due to the organic capping molecules.

Since conducting layers are an important requirement for applications, a lot of research


focuses on ways to improve the conductivity. One solution is to heat the superlattice after

it has formed. At a high enough temperature, the ligands evaporate and the distance

between nanocrystals decreases. Unfortunately, the temperatures required to remove com-

monly used ligands also lead to sintering of the crystals and thereby losing the confinement

properties. This has been solved by exchanging oleate ligands with a ligand that degrades

at a temperature where sintering does not yet occur [33].

An alternative route, without heating, is exchanging the ligands with much smaller ones by

treating the superlattices with an appropriate reagent. These ligands can be, for example,

carboxylic acids with shorter carbon chains, or inorganic anions like S2-, OH-, NH2-, and

BF4- [31,16,17,24,27].

Instead of carrying out a ligand exchange reaction at the QD surface, it is also possible

to exploit the free energy difference between different crystal facets. Choi et al. used

DFT methods to calculate binding energies of the acetate molecule on different PbS QD

facets and found that it binds more strongly on the (111) facets than on the (100) facets,

with a difference of 0.346 ± 0.029 eV [5]. Baumgardner et al. used this to investigate

facet specific adsorption/desorption equilibria of oleate molecules on PbSe nanocrystals.

By treating self-assembled hexagonal superlattices with solvents that dissolve oleic acid,

preferential desorption at the (100) facets was observed, resulting in oriented attachment

along this direction. The conductivity of these layers increased by more than 3 orders of

magnitude [2,13].

All these post-treatments, however, lead to an inherent volume decrease in the superlattice,

and as a consequence unwanted crack formation occurs. Recently, Simon et al. showed that

low temperature annealing over a period of 24 h leads to an increased formation of small,

crystalline linkages between the nanocrystals [29]. This could prove an excellent method to

increase conductivity of the superlattices, while still preserving confinement. At the same

time, the volume decrease is minimal because the ligand shell remains mostly intact, and

this could strongly decrease crack formation.

Another way to prevent cracks is by exchanging the ligands before forming a superlattice.

Instead of treating the superlattice with a reagent, the nanocrystals are treated in solution.


Exchange with small inorganic ligands like S2- or OH- is typically carried out in a two-phase

system, where the apolar phase contains the nanocrystals capped with organic ligands and

the polar phase contains a salt of the exchanging ion. A successful exchange is signaled

by a phase transition of the crystals to the polar phase, usually formamide or dimethyl

sulfoxide (DMSO). This method has been successfully applied to a number of different

nanocrystals (e.g. CdSe, ZnS, InP, F2O3, Au) with a number of different inorganic ligands

(e.g. S2-, OH-, NH2-, BF4-) [23,6]. However, due to problematic charge interactions between

the crystals, a stable dispersion of inorganically capped lead chalcogenide nanocrystals has

not been reported yet.

New, promising methods to obtain conducting superlattices are a combination of pre- and

post-treatment, and influence the core-shell chemistry at the moment of superlattice for-

mation. For these methods to work, a solution of nanocrystals is cast upon a non-solvent,

where the crystals are allowed to form a superlattice. In this way, the nanocrystals still

preserve enough mobility to allow some movement. The relative higher mobility, compared

to nanocrystals on a solid substrate, enables the superlattice to contract uniformly with-

out forming large cracks. The choice of the liquid subphase and/or additives provides the

chemist with some useful parameters. Dong et al. assembled FePt, Au, and PbS superlat-

tices on an acetonitrile subphase, and carried out an in-situ ligand exchange by injecting

an appropriate reagent into the subphase after initial superlattice formation. The macro-

scopic contraction of the floating nanocrystals avoided crack formation and preserved the

nanocrystal ordering, significantly improving electrical transport in the superlattices [7].

Evers et al. reported PbSe superlattices formed by oriented attachment along specific

crystal facets, depending on the concentration of the nanocrystal solution. Ethylene gly-

col was used as subphase, and they showed that by keeping it at elevated temperatures,

oleate ligands absorb in the subphase, making the crystal facets available for interparticle

bonding [10].


1.3 Application: photodetectors

Photodetection is a broad field of research and it is almost impossible to underestimate

its technological importance. The electromagnetic spectrum, shown in Figure 1.4, encom-

passes a wide range of energy and each region is a source for a multitude of applications.

Photodetectors are typically used in the region from far infrared up to gamma rays and

applications can be grouped in two main types; communication and remote sensing. For

communication purposes, the radiation is used as a carrier for an encoded signal, while

for remote sensing applications, the radiation itself is the signal, and contains information

about an object or a substance. To give a faint idea about the range of applications, pho-

todetectors are used in motion detectors, all types of cameras, from normal video camera

to night vision, pollution detection in individual cells up to space-based environmental

monitoring, fiber-optic communication systems, detectors used in telescopes etc. The type

of application determines the specific technical requirements, such as the spectral range in

which the detector is sensitive, the speed at which successive signals can be distinguished,

the maximum optical power it can handle, the working temperature, size, robustness, and

last but not least, the production cost.

Figure 1.4: A presentation showing the different regions of the electromagnetic spectrum with

their corresponding wavelength and frequency.

Basically, a photodetector converts incident photons to electrical signals. An incident

photon can be absorbed by an electron, and depending on the energy of the photon, the


electron is either completely freed from its atomic or molecular environment, i.e. exter-

nal photoelectric effect, or gets excited to a higher atomic or molecular energy level, i.e.

internal photoelectric effect. External photoelectrons are typically observed in the X-ray

part of the spectrum, and are not considered here. Formation of internal photoelectrons

on the other hand, requires much lower energies and can be observed in the UV, visible,

and infrared part of the spectrum.

Photodetectors based on the internal photoelectric effect can be divided in two main classes:

photoconductors and photodiodes.

1.3.1 Photoconductors

A photoconductor typically consists of a semiconducting material between two biased metal

contacts. When an incident photon with sufficient energy is absorbed, an electron is excited

to the conduction band. The created electron-hole pair exists for a certain time until the

electron relaxes back to its ground state in the valence band. As long as the electron

remains in an excited state, a charge carrier circulates through the external electrical

circuit, creating a photocurrent in the semiconducting material. When the charge carrier

circulates the circuit multiple times while the exciton exists, a condition of photoconductive

gain is created. Trap states, which are intermediate energy levels originating from surface

defects or molecules bound to the QD surface, effectively increase the amount of time an

electron remains in an excited state, and thus the time a charge carrier exists. This leads

to an expression for the photoconductive gain G as

G =τlifetime

τtransit(1.9)

where τlifetime is the lifetime of the exciton and τtransit the time it takes for a free charge

carrier to pass from one metal contact to the other. A gain greater than 1 is thus obtained

when the exciton lifetime exceeds the time it takes for the charge carrier to travel between

electrodes. The mechanism of photoconductivity is schematically illustrated in Figure 1.5.

The internal signal amplification provided by photoconductive gain enables high sensitivi-

ties, since one created charge carrier creates a photocurrent that is many times higher. At


the same time, construction of electronic circuits is simplified because external amplifica-

tion is not needed if the gain is sufficiently high. However, the long exciton lifetime (∼ µs)

makes the detector respond relatively slowly to a time-changing photon flux, which limits

the range of frequencies at which these highly sensitive photoconductors can be used. This

trade-off between photoconductive gain and response time has an impact on the type of ap-

plications it can be used for. A photoconductor with high gain would, for example, not be

suited for a communication type of application, since these usually employ high frequency

signals in the GHz range. However, it is appropriate to use high gain photoconductors for

gas sensing, in order to detect e.g. toxic compounds at very low concentrations.

Figure 1.5: Schematic representation of the mechanism of a photoconductor. After absorption

of a photon, an electron is excited to the conduction band (CB). The created hole in

the valence band (VB) generates a photocurrent by circulating through the external

circuit. Trapping of the electron in a trap state (TS) increases the lifetime of the

exciton and thereby increases the photocurrent. After relaxation to the ground

state, the photocurrent falls back to 0.

1.3.2 Photodiodes

Photodiodes are very commonly used in today’s technology. They make use of the internal

electrical field of a pn-junction to propel photogenerated holes and electrons in opposite

directions. The separated charges are detected by a voltage difference across the junction.

Photodiodes typically have a very fast response time, making them suitable for high fre-


quency applications. On the downside, photodiodes can’t produce gain. This type will not

be discussed further, as it reaches outside the scope of this thesis.

1.4 Photoconductors based on quantum dot superlat-

tices

The subject of this thesis is to investigate the use of QDs as photosensitive material in

a photoconductor type of detector. Colloidal QDs have the advantage of being dispersed

in solution, thus enabling solution-based deposition techniques such as dropcasting, dip-

coating, Langmuir-Blodgett etc. Compared to conventional vacuum deposition techniques

such as chemical vapor deposition (CVD), sputtering, or molecular beam epitaxy (MBE),

solution-based techniques are low cost, easy to perform and fast. An extra benefit is that

the size of the QD allows spectral tuning of the band gap via the quantum confinement

effect. The downside of colloidal QDs is that their intrinsic absorption coefficient is lower

than the bulk material due to the local field factor. This means a less efficient photon-to-

current conversion.

The QD superlattice has to meet three important requirements in order to be useful for

photodetection. The first one is that the superlattice is highly conductive, since it allows

large photocurrents to travel through the layer. This is envisioned by a superlattice con-

sisting of electronically coupled QDs, providing high charge carrier mobilities. A second

requirement is that the superlattice has a low intrinsic carrier concentration. This reduces

the dark current and results in a higher contrast with the photocurrent, thereby enabling

higher sensitivities. The third requirement is that the superlattice is homogeneous and

crack-free over a large area, so that reliable and reproducible results can be obtained. An

possible extra benefit of a large area connected superlattice is that it increases the local

field factor and thus the intrinsic absorption coefficient of the material, since this structure

is somewhere between individual QDs and bulk material.

Chapter 2

Quantum Dot Synthesis

A first step towards the fabrication of photodetectors is the bottom-up synthesis of colloidal

QDs. This chapter first describes the synthesis of highly monodisperse PbSe nanocrystals

and their subsequent characterization by XRD, TEM, and NIR absorbance spectroscopy.

As an aid for the next chapter, this chapter also shows how high resolution TEM images

are used to obtain information on the crystallographic orientation of the QDs.

2.1 Synthesis of PbSe QDs

PbSe semiconductor nanocrystals were synthesized by the hot injection method. Two pre-

cursors were prepared separately before the synthesis was carried out. The first precursor,

lead oleate, was prepared by mixing 3.74 g PbO with 15.87 mL of oleic acid at 150 ◦C

for one hour. To the still hot, clear solution, 67.96 mL of diphenyl ether was added. The

Selenium precursor, TOPSe, was prepared by dissolving 3.52 g Se in 46.59 mL tri-octyl

phosphine (TOP) and 0.41 mL diphenyl phosphine at 150 ◦C. From the lead precursor,

20.5 mL was taken and heated to 180 ◦C. At this point, 15 mL of TOPSe was rapidly

injected. Successful nucleation was indicated by a blackening of the solution. The crystals

were further grown at 150 ◦C for 60 s, after which the reaction was quenched with 15 mL

butanol. The resulting nanocrystals were precipitated with 10 mL acetonitrile, a miscible

15

Chapter 2. Quantum Dot Synthesis 16

non-solvent, and centrifuged for 2 minutes at 10000 RPM. The liquid phase was decanted

and the quantum dots were again dispersed in 6 mL toluene. After a second wash with ace-

tonitrile and toluene the purified quantum dots were stored in a nitrogen purged glovebox.

All procedures were carried out under inert atmosphere.

2.2 Results and discussion

The TEM image in Figure 2.1A shows that the hot injection synthesis yielded monodis-

perse, quasi-spherical PbSe nanocrystals. The absorbance spectrum of the colloidal so-

lution is shown in Figure 2.1B. Three exciton peaks are seen at 1716 nm, 1350 nm, and

1034 nm, which are respectively transitions between the S-S, S-P, and P-P electronic states

of the PbSe QDs. The first exciton peak at 1716 nm was used to determine the average

QD diameter and size distribution via a commonly used method [21]. This resulted in an

average nanocrystal diameter of 6.0 nm ± 4.25 %. The method was also used to determine

the concentration of the stock solution (47.7 µM).

Figure 2.1: A: A TEM image of the as-synthesized PbSe QDs showing a quasi-spherical shape

and monodisperse size distribution. B: The absorption spectrum of a 200 times

diluted solution corresponding to the QDs in Figure A.

The XRD pattern in figure 2.2 confirms that the PbSe nanocrystals have a cubic crystal

structure, as is the case with bulk PbSe. The positions of the peaks were used to calculate


the interplanar distances by application of Bragg’s law (Equation 2.1), where n is the

diffraction order, λ the wavelength of the X-ray, dhkl the interplanar distance and θ the

angle between the incident ray and the scattering plane. The results are shown in Table

2.1. It should be noted that the sharp peaks in the XRD pattern originate from diffraction

of the Si substrate.

Figure 2.2: XRD pattern of the as-synthesized PbSe QDs showing diffraction peaks at the same

positions as cubic bulk PbSe.

nλ = 2dhklsinθ (2.1)

2θ (◦) dhkl (A) hkl

25.15 3.54 111

29.12 3.06 200

41.65 2.17 220

Table 2.1: Interplanar distances and indices of the planes, as determined from the peaks in the

XRD pattern.


High resolution images were used to measure the angles between crystal facets and the

distance between atomic planes. Table 2.2 summarizes the measured angles and distances

of some typical orientations of nanocrystals, shown in Figure 2.3. The angles are averaged

over the 8 angles around the nanocrystal, the distance is an average of 100 measure points.

The measured values, combined with the calculated interplanar distances from the XRD

pattern, allow a straightforward identification of the observed planes. Figure 2.3 uses a

crystal model consisting of 6 (100) planes and 12 (110) planes to visualize the orientation of

the nanocrystal on the TEM grid (A). When individual atoms are seen in a cubic structure,

as in Figure 2.3B, with an interatomic distance close to 3.06 A and angles between the

facets around 135◦, the particle is looked upon from the (100) direction and this plane is

parallel to the substrate. In the case that only atomic rows are seen, as in Figure 2.3C and

D, the particles are rotated around a specific axis, which is determined from the distance

between the atomic rows. A distance close to 3.06 A means that the particle is rotated

around a (100) axis and when a distance close to 2.17 A is found, the particle is rotated

around a (110) axis, as illustrated by the crystal models.


Figure 2.3: A: A PbSe unit cell and a crystal model of a 6 nm QD consisting of 6 (100) planes

and 12 (110) planes. B-D: TEM images and corresponding crystal models indicating

the lattice planes in several orientations of QDs. The particles are either not rotated

(B), rotated along a (100) axis (C), or rotated along a (110) axis (D).

Particle Distance atomic rows (A) Facet angles (◦)

Figure 2.3B 3.02 ± 0.05 135 ± 1.6

Figure 2.3C 3.05 ± 0.02 134 ± 3.6

Figure 2.3D 2.16 ± 0.10 134 ± 1.5

Table 2.2: The interatomic distances and angles between facets as measured from the corre-

sponding TEM images in Figure 2.3.

Chapter 3

Quantum Dot Superlattices

3.1 Superlattices by dropcasting on ethylene glycol

A first straightforward method to produce QD superlattices is by simply dropcasting a

QD solution. The use of a liquid subphase, onto which the QD solution is cast, not

only allows an easy transfer of the formed superlattice to other substrates, it also provides

opportunities to influence and control the formation of the superlattice. This section shows

how the composition of the subphase, the temperature of the subphase, and the rate at

which the QDs are added influences the microscopic morphology of the superlattice.

To form the superlattices shown in this section, 10-50 µL of a PbSe QD solution with

concentration varying from 0.29 µM to 2.9 µM, was cast onto 1 mL ethylene glycol

(HO− C2H4 −OH), in a glass container with 10 mm diameter. Upon addition of the

solution, the QDs spread on the liquid surface as the solvent evaporated. This left a thin

film on the subphase, and samples of the superlattice were taken at the center of the vial

after 1 h, either by fishing on a TEM grid or stamping on a Si or glass substrate. After de-

position, the substrate was immersed in deionized water to remove residual ethylene glycol.

The sample was then dried under a gentle nitrogen flow. The glass vial and substrates were

cleaned beforehand by rinsing with aceton, iso-propanol, and water. To prevent oxidation

of the QDs, all dropcasting experiments were carried out in a nitrogen purged glovebox.

20

Chapter 3. Quantum Dot Superlattices 21

3.1.1 Addition of Na2S to the subphase

Figure 3.1 shows the microstructure of monolayers formed by dropcasting 18 µL of a 0.29

µM QD solution at room temperature. This is the amount of nanocrystals needed to

form one monolayer of a cubic superlattice. From left to right, the amount of Na2S in the

subphase is increased from naught to a 100 fold excess, compared to the total amount of

ligands in the QD solution. This was calculated for 6 nm nanocrystals with an average

ligand density of 4 ligands/nm2 [14].

The superlattice formed on pure ethylene glycol (Figure 3.1A) shows a hexagonal distri-

bution. The average distance between the nanocrystals is 2.9 ± 0.8 nm, corresponding to

an intertwingled ligand shell stabilized by Van der Waals interactions (the length of oleic

acid is 1.97 nm). This indicates that the nanocrystals are surrounded by a complete ligand

shell, and that ethylene glycol at room temperature does not affect the ligand shell at the

QD surface. The surface density of nanocrystals in this layer is 15.56 · 103 QDs/µm2, as

determined from the TEM image.

The addition of a 10 fold excess Na2S to the subphase (Figure 3.1B) results in a fraction

of the nanocrystals connecting to each other, with an accompanying deviation from the

hexagonal structure. Although a significant fraction of nanocrystals remains separated

by the organic capping molecules, the increase in interparticle connections, compared to

no additional Na2S, is clear. Adding more Na2S subsequently increases the amount of

interparticle connections, as shown by the layer formed with a 50 fold excess in Figure

3.1C. At the same time, the surface density of nanocrystals increases due to a more closely

packed structure. Eventually, with even more Na2S in the subphase, the density of the

layer does not exhibit a dramatic increase, since at 50 fold excess most of the nanocrystals

are already connected. Figure 3.1D shows the layer formed on a subphase containing a 100

fold excess Na2S. The surface density, as determined from the TEM image, is 21.86 · 103

QDs/µm2.


Figure 3.1: TEM images showing the microstructure of superlattices formed by dropcasting a

QD solution on ethylene glycol. From left to right the amount of Na2S added to

the subphase is no Na2S (A), a 10 fold excess (B), a 50 fold excess (C) and a 100

fold excess (D). Increasing amounts of Na2S increases the amount of interparticle

connections and the surface density in the superlattice. Scale bars are 10 nm.

Figure 3.2 shows more detailed TEM images of the connections between nanocrystals. The

distance between the lattice planes in Figure 3.2A is 3.01 ± 0.16 A, which is approximately

the distance between Pb and Se along a (100) axis. The crystals are thus rotated along

this axis, and bind via (100) planes. Another example is shown in Figure 3.2B where the

interplanar distance is 2.18 ± 0.14 A. This is approximately the distance between (110)

planes, meaning that the nanocrystals are rotated along the (110) axis. This figure shows

that, although the orientation is different, the bonding still occurs via the (100) planes.

From inspection of all recorded TEM images, all connections between crystals were found

to occur via (100) planes, no bonding via (110) or (111) planes was observed.


Figure 3.2: High resolution TEM images showing interparticle connections via (100) planes

where the nanocrystals are slightly rotated along a (100) axis (A) or a (110) axis

(B).

The FTIR spectrum in Figure 3.3 shows how the addition of Na2S to the subphase affects

the ligand shell of the QDs. Without any Na2S, the spectrum shows relative strong C− H

stretch bands, originating from the oleate ligand shell. At a 50 fold excess, the bands

are strongly decreased in intensity, and at a 100 fold excess, no significant C− H stretch

bands are distinguishable. This shows that Na2S effectively removes oleate ions from the

QD surface, allowing the QDs to form interparticle connections. This is in agreement with

what is seen in the TEM images.


Figure 3.3: FTIR spectra showing the C−H stretch vibrations of superlattices formed on ethy-

lene glycol containing no Na2S (green), a 50 fold excess (blue) and a 100 fold excess

(red). The signals decrease with increasing amounts of Na2S, indicating removal of

the native oleate ligands.

Figure 3.4 shows SEM images of the layers formed on ethylene glycol containing a 10 fold

and a 50 fold excess Na2S (A and B respectively). A very homogeneous monolayer is

deposited on the Si wafer and the thickness is uniform over an area up to tens of µm2.

Figure 3.4: SEM images of superlattices formed on ethylene glycol containing a 10 fold excess

Na2S (A) and a 50 fold excess (B). Both superlattices consist of a homogeneous

monolayer with very few cracks over a large area (scale bars are 1 µm).


3.1.2 Temperature of the subphase

The effect of temperature on the formation of QD superlattices was studied by heating the

subphase to 50 ◦C before adding the QD solution. However, when monolayer amounts were

added, no layer was deposited on the substrate, most likely because convection currents in

the subphase drifted the nanocrystals away from the center of the vial. For this reason,

more material was added to the subphase.

Figure 3.5 shows the effect of adding 50 µL of a 1.4 µM QD solution to ethylene glycol at

room temperature (A) and at 50 ◦C (B). At room temperature, the nanocrystals arrange in

a hexagonal close packed structure and the distance between the crystals indicate that the

ligand shell has not been altered in a significant way. On the other hand, at elevated tem-

perature, the individual nanocrystals merge together and form a superlattice of connected

QDs with a quasi-cubic symmetry. The superlattice has a density of 22.13 · 103 QDs/µm2.

Figure 3.6 shows more detailed images of some connections found in the supperlattice.

The nanocrystals in Figure 3.6A are viewed along the (100) direction, as determined from

measurements of the angles (135 ± 2◦) and interatomic distances (3.04 ± 0.19 A) of crystal

1. This orientation means that the nanocrystals fuse via their (100) facets. In Figure 3.6B,

crystal 1 has an average interatomic distance of 3.05 ± 0.18 A and an average angle between

the facets of 135 ± 1◦, crystal 2 has an average distance of 3.05 ± 0.15 A between the

atomic rows. As indicated in the figure, this means that the bond between crystal 1 and

2 happens via the (100) facets. The bond between crystal 2 and 3 can either happens

via (100) or (110) facets, depending on the rotation of the nanocrystal. However, bonding

via the (110) facets means that the (100) facets between crystal 1 and 2 are rotated 45◦

relative to each other. Due to a large lattice mismatch between these orientations, this type

of bonding is very unlikely, which suggests that the bond between crystal 2 and 3 happens

via the (100) facets. For the same reason, bonding between different facets, between (100)

and (110) for example, is rather unlikely. For clarity, the arrangement of atoms in the

different possible orientations are shown in Figure 3.7.

From careful inspection of TEM images, it is found that the majority of connections happen


via (100) facets. Although fusion via (110) facets is also possible, this is not frequently

observed. An example of (110) bonding is shown in Figure 3.6C, where the average distance

between the atomic rows is 3.04 ± 0.16 A.

Figure 3.5: TEM images of the superlattices formed on pure ethylene glycol at room temper-

ature (A) and at 50 ◦C (B). The increased temperature leads to a more dense

structure of interconnected QDs, compared to a hexagonal close packed layer with

a larger interparticle distance at room temperature. Scale bars are 50 nm.

Figure 3.6: High resolution TEM images of interparticle connections found in the superlattice

formed at 50 ◦C. Most connection occur via (100) planes (A and B), although

bonding via (110) is also possible (C).


Figure 3.7: Crystal models showing the arrangement of Pb2+ and Se2- ions at the (100) and

(110) planes.

The FTIR spectrum in Figure 3.8 compares the layer formed at room temperature (green

curve) with the layer formed at 50 ◦C (blue curve). In the layer formed at room tem-

perature, strong absorption bands from C− H stretch vibrations are seen at 2925 cm-1

and 2850 cm-1. At 50 ◦C, a strong reduction of these signals is seen, indicating that the

majority of the ligand shell is removed from the QD surface. This shows that not only

the addition of a reagent to the subphase, but also moderately increasing the temperature

of the subphase is a successful strategy for removal of the ligand shell, and subsequently

leads to the formation of interparticle connections.


Figure 3.8: FTIR spectra showing the C−H stretch vibrations of superlattices formed on pure

ethylene glycol at room temperature (green) and at 50 ◦C (blue). The strong reduc-

tion of the signals indicate that the majority of native oleate ligands are removed.

The topography of the layer was measured with AFM. According to the surface profile in

Figure 3.9A, which follows the line drawn in Figure 3.9B, the superlattice mainly consists

of mono- and bilayers. Since more QDs were added to the subphase, this is a reasonable

result. Furthermore, the topography image shows that the layer contains small cracks with

widths around 1 µm (black lines), and that aggregations of up to 100 nm are spread across

the surface (yellow to white lines). This morphology is confirmed by the SEM images in

Figure 3.10. Figure 3.10A shows more in detail the cracks across the surface and Figure

3.10B is a higher magnification image showing the presence of mono- and bilayers. It

is clear that these layers are less homogeneous than the ones obtained in the previous

paragraph.


Figure 3.9: Surface profile (A) following the line drawn in the topography image (B) of the

superlattice formed on pure ethylene glycol at 50 ◦C. The superlattice primarily

consists of a bilayered structure with small portions consisting of a monolayer.

Figure 3.10: SEM images of the superlattice formed on pure ethylene glycol at 50 ◦C. The

images confirm the presence of cracks (A - scale bar 1 µm) and varying regions of

mono- and bilayers (B - scale bar 100 nm) as seen in the AFM image.

3.1.3 Rate of QD addition

Another parameter that influences the superlattice formation is the rate at which the

nanocrystals are added to the subphase. Figure 3.11 shows the different morphologies

obtained by adding 50 µL of a 0.9 µM QD solution either in one swift motion (A), or


in a drop-wise fashion at a rate of about 1 drop/s (B). In both cases, the solutions were

added at 50 ◦C. This section shows the results of adding a 0.9 µM QD solution in order

to properly compare with the structures formed at room temperature (only results with a

maximal concentration of 0.9 µM available). However, the molten structures were obtained

at different temperatures and with different concentrations, and addition of 1.4 µM resulted

in an identical morphology as the one shown here.

The fast addition results in superlattices where individual nanocrystals connect together,

similar to the superlattices obtained in the previous paragraph. On the other hand, when

the nanocrystals were added drop-wise, the morphology drastically changed. The crystals

fuse together and form a molten-like network where the individual nanocrystals cannot

be distinguished anymore. At the same time, the surface density of material decreases

considerably.

Figure 3.11: TEM images showing the microstructure of superlattices formed on pure ethylene

glycol at 50 ◦C, either by fast addition (A) or by drop-wise addition of the QD

solution (B). A drop-wise addition drastically changes the morphology from a

connected layer of QDs to a molten-like structure. Both scale bars are 50 nm.

The morphology of the layer at elevated temperature differs from that formed by drop-

wise addition at room temperature (Figure 3.12). Here, the nanocrystals arrange in a very

periodic structure over a large area. This is indicated by the Fourier transform, showing

well defined points in the reciprocal lattice (inset Figure 3.12A). From Figure 3.12B it


is clear that these structures are a stacking of hexagonal close packed layers where the

crystals are separated by their ligand shell. In general, the layers have a thicker stacking

of monolayers, compared to a fast addition at room temperature.

Figure 3.12: A: Superlattice formed by drop-wise addition of the QD solution at room tempera-

ture. The Fourier transform in the inset shows a long range hexagonal periodicity.

Scale bar is 50 nm. B: A higher magnification image of the same structure showing

that the superlattice consists of a stacking of hexagonal close packed layers. Scale

bar is 10 nm.

The FTIR spectrum in Figure 3.13 confirms that at room temperature the ligand shell

remains intact (green curve), as was the case in the previous section. Again, adding the

nanocrystals to the subphase at 50 ◦C removes the ligands from the QDs, as indicated by

the loss of C− H stretch vibrations in the spectrum (red curve). This means that removal

of the ligands is not significantly influenced by the rate at which the nanocrystals are added

but that it is, in this case, mainly determined by the temperature of the subphase.


Figure 3.13: FTIR spectra showing the C−H stretch vibrations of superlattices formed by

drop-wise addition of the QD solution at room temperature (green) and at 50

◦C (red). The strong reduction of the signals show that almost all native oleate

ligands are removed.

AFM was used to determine the thickness of the layer. Figure 3.14A shows the surface

profile following the line drawn in Figure 3.14B. The layer has a thickness of about 11 nm,

but since this layer is not comprised of individual nanocrystals, nothing can be said about

the thickness in terms of mono- or bilayers.


Figure 3.14: Surface profile (A) following the line drawn in the topography image (B) of the

superlattice formed by drop-wise addition of the QD solution at 50 ◦C. The profile

shows a thickness of about 11 nm.

In general, the long-range morphology of this layer is not as good as for the layers shown

previously. Rather than being uniform, the layer consists of a patchwork of smaller chunks

or flakes of material, with sizes of several µm2. This is illustrated by the SEM image in

Figure 3.15.

Figure 3.15: SEM image showing a small piece of the superlattice formed by drop-wise addition.

Scale bar is 1 µm.


3.2 Superlattices by Langmuir-Schaefer deposition

QD superlattices were also made by the formation of Langmuir films (see Appendix A for

a short explanation of the technique). For this, 100 µL of a filtered 2.46 µM QD solution

in hexane was added drop-wise to the Langmuir trough. After the solvent evaporated, the

nanocrystals were slowly compressed by two barriers, reducing the available surface area by

10 cm2/min. After compression to the desired pressure, samples were taken by stamping

the layer on a TEM grid, or a Si or glass substrate (i.e., Langmuir-Schaefer deposition).

The TEM images in Figure 3.16 show PbSe QD monolayers formed by Langmuir-Schaefer

deposition at room temperature. From left to right, the subphase was varied from water

to ethylene glycol to diethylene glycol. On water, the QDs arrange in a typical hexagonal

close packed structure, stabilized by Van der Waals interactions between neighboring ligand

shells. A similar ordering is seen with ethylene glycol as subphase. The interparticle

distance is comparable to the layer on water, indicating that at an elevated pressure of 25

mN/m, ethylene glycol does not significantly affect the ligand shell. However, the mobility

of the crystals on ethylene glycol is a bit lower due to stronger Van der Waals interactions

between ethylene glycol and oleate, and at 25 mN/m, the layer is a little less homogeneous

than on water (macroscopically).


Figure 3.16: TEM images of the superlattices formed by Langmuir-Schaefer deposition on wa-

ter (A), ethylene glycol (B) and diethylene glycol (C). (A) and (B) were taken at a

pressure of 25 mN/m and show a hexagonal close packed structure with no inter-

particle connections. (C) was taken at 4 mN/m and shows a disordered structure

of interconnected QDs. Scale bars are 10 nm.

On diethylene glycol, the morphology of the layer drastically changes from a dense hexag-

onal structure, to a less dense, somewhat disordered structure of interconnected nanocrys-

tals. This suggests that diethylene glycol, as opposed to ethylene glycol, effectively removes

the ligand shell at room temperature. Figure 3.17A shows a more detailed image of some

interconnected nanocrystals. The average distance between the atomic planes in Figure

3.17B is 3.02 ± 0.17 A, determined from 100 measure points. This shows that the crys-

tals are rotated around a (100) axis, and that connections in the direction perpendicular

to the lattice planes occur via (100) planes. Connections in the other direction, parallel

to the lattice planes, either occur via (100) or (110) planes, but as explained earlier, the

most probable bonding occurs via (100) planes due to lattice mismatches. A model of the

connected QDs with and without crystal shape is shown in Figure 3.17C.

The irreversible bonding between crystal facets introduces a rigidness to the structure, in

the sense that once bonds are formed in a small group of nanocrystals, they cannot easily

rearrange in a more close packed structure. Upon compression, this eventually leads to

aggregations of QDs and the layer gets an overall inhomogeneous thickness and morphology.


For this reason, the layer on diethylene glycol was sampled at 4 mN/m, instead of at 25

mN/m, as was the case with water and ethylene glycol.

Figure 3.17: A: High resolution TEM image of the structure formed on diethylene glycol. B:

High resolution TEM image showing that the connection between nanocrystals

occur via (100) planes and that they are slightly rotated along a (100) axis. C:

Crystal models of the bonds shown in (B) with and without crystal shape.

The formation of interparticle connections on diethylene glycol is in line with the desired

result, although the strong inhomogeneity is problematic for use in applications. In an

attempt to better control the formation of connections, monolayers were formed on a

subphase consisting of 70 % ethylene glycol and 30 % diethylene glycol. On this subphase,

the nanocrystals spread better than on pure diethylene glycol, and it was possible to

compress the layer to 12 mN/m without formation of problematic aggregations. After

compression, the layer was left to stabilize for 1 hour, during which the layer contracted,

and an accompanying drop in pressure was seen, from 12 mN/m to 9.4 mN/m. The

contraction was primarily in the direction parallel to the compressing bars of the apparatus,

and further compression only led to inhomogeneities in the layer.

The layer stamped on a Si wafer has a very homogeneous appearance across the surface.

This is confirmed by the SEM image in Figure 3.18A, which shows that the layer is homo-

geneous over an area of several tens of square micrometer. Some small ridges are visible,


originating from the applied compression after the layer had contracted. Figure 3.18B

shows that the surface is not completely covered with nanocrystals and thus the density of

the layer is not optimal, but overall, the QDs are evenly spread across the surface.

Figure 3.18: SEM images of the superlattice formed by Langmuir-Schaefer deposition on a sub-

phase consisting of 70 % ethylene glycol and 30 % diethylene glycol. A relatively

homogeneous film over a large area is obtained (A - scale bar 10 µm) although

the higher magnification image shows that the density of QDs is not maximal (B

- scale bar 100 nm).

The TEM image in Figure 3.19A confirms that the layer is not as dense as possible and that

many open gaps are present. A significant amount of QDs show interparticle connections

but at the same time, many nanocrystals retain their native ligand shell. The obtained su-

perlattice thus becomes a combination of connected and unconnected QDs (Figure 3.19B).

As expected from previous results, most connections occur via (100) planes. This result

confirms that DEG is a very good removing agent for the ligand shell, and that the amount

of DEG in the subphase determines the fraction of ligands that are lost.


Figure 3.19: TEM images showing the microstructure of the superlattice formed by Langmuir-

Schaefer deposition on a subphase consisting of 70 % ethylene glycol and 30 %

diethylene glycol. The scale bar in (A) is 100 nm, in (B) 10 nm. (C) is a high

resolution TEM image showing interparticle connections via (100) planes.

Inspired by the results in section 3.1.1, monolayers were formed on ethylene glycol con-

taining a 100 fold excess Na2S. The addition of Na2S changed the subphase in such a

way that the spread of the nanocrystals was retarded, leading to small spots of material

(approximately 5 cm in diameter) at the place where the drop of QD solution was added.

This decrease of QD mobility indicates an increased interaction with the subphase. After

compression to 12 mN/m, the layer was left to stabilize for 1 hour and again, a contraction

of the layer was seen. Due to the decreased mobility, aggregations were formed at the place

where the spots collided, although these were manageable and still homogeneous areas of

several square centimeters were obtained.

The TEM image in Figure 3.20A shows that a relatively dense layer is obtained where the

QDs have a rather random, non-symmetrical stacking. A large fraction of the QDs show

interparticle connections (Figure 3.20B), but also many QDs are not connected, signaling

an incomplete ligand removal.


Figure 3.20: TEM images of the superlattice formed by Langmuir-Schaefer deposition on an

ethylene glycol subphase containing a 100 fold excess Na2S.

Figure 3.21 shows the SEM image of a layer stamped on a Si wafer. The layer is somewhat

less homogeneous, with overall dense and less dense areas. However, no cracks or large

aggregations are seen across the surface.

Figure 3.21: SEM images of the superlattice formed by Langmuir-Schaefer deposition on an

ethylene glycol subphase containing a 100 fold excess Na2S.

Figure 3.22 shows the FTIR spectrum of the layers formed by Langmuir-Schaefer deposi-

tion. The green curve represents the spectrum of a monolayer formed on a H2O subphase.

In this layer, all organic ligands remain on the QD surface, and strong C− H stretch vi-

brations are seen. Both the addition of DEG (red curve) and the addition of Na2S (blue


curve) to the subphase result in a decrease of the C− H stretch bands. This indicates that

both additions have an effect on the ligand shell as partial removal of the organic ligands

is seen. Although a stoichiometric 100 fold excess of Na2S is added, ligand removal is not

complete, as was the case in section 3.1.1. This can be attributed to the larger volume

of the subphase, effectively decreasing the concentration of Na2S, and thus reducing the

available S2- ions at the subphase-QD interface.

Figure 3.22: FTIR spectra of the Langmuir-Schaefer films formed on different subphases: water

(green), ethylene glycol containing a 100 fold excess Na2S (blue) and ethylene

glycol containing 30 % diethylene glycol (red). Both additions to ethylene glycol

reduce the amount of oleate ligands, although a significant amount is still present.

3.3 Light absorption by superlattices

Figure 3.23 shows the absorption spectra of two superlattices. These structures were

selected because they illustrate the effect of interparticle connections and because they were

both very homogeneously deposited on a glass substrate. The red curve is the spectrum

of a hexagonal monolayer without interparticle connections, formed by Langmuir-Schaefer

deposition on water (see Figure 3.16A). The blue curve is the spectrum of the connected

superlattice formed by dropcasting on ethylene glycol containing a 100 fold excess Na2S


(see Figure 3.1D). The absorption onset of the connected superlattice is somewhat shifted

to longer wavelengths, followed by a steeper slope towards shorter wavelengths, resulting

in a higher absorbance at high photon energies.

The absorption at 400 nm and the density of the layer (Ns) were used to calculate a cross

section for the PbSe nanocrystals in both superlattices, by neglecting the reflection factor

in Equation 3.1. The result is summarized in Table 3.1. The density was determined from

the corresponding TEM images.

Figure 3.23: Absorbance spectra comparing the absorbance of an unconnected superlattice

(red) with a connected superlattice (blue).

σ = ln10A−RNs

(3.1)

A400nm Ns(10−3QDs/nm2) σ(10−14cm2)

Unconnected superlattice 0.11462 19.85 13.30

Connected superlattice 0.32683 21.86 34.42

Table 3.1: Cross sections of the connected and unconnected superlattice, calculated from Equa-

tion 3.1 without taking into account reflection of the superlattices.


3.4 Discussion

3.4.1 Oleate and the subphase

The organic ligand, oleate, is a surfactant molecule with a hydrophobic C18 carbon chain

and a hydrophilic carboxylate head group. The ligand is bound to Pb2+ ions at the QD

surface with the carboxylate group, forming Pb-oleate [14]. The high density of ligand

molecules, approximately 3-4 ligands/nm2, gives the faceted nanocrystal a quasi-spherical

shape. In other words, when the solvent is evaporated and the nanocrystals are allowed

to self-assemble, the ligand shell is responsible for the hexagonal close packed structure,

being the most dense and thus energetically favored stacking of an assembly of spheres. As

the ligands are removed from the QD surface, the non-spherical shape of the nanocrystal

becomes more expressed and the connected QD superlattice adopts a structure which

depends on the crystal structure and the crystal shape of the QD building blocks. In the

case of PbSe QDs, a tendency to form a cubic structure (although not perfect) is seen.

The experiments in this chapter show that several strategies can be used to remove the

organic capping molecules in order to form connected superlattices. A first route is the

addition of Na2S to the subphase. This removes the ligands by reacting with Pb-oleate

to form PbS and Na-oleate (Reaction 3.2). In this way, the excess Pb2+ ions at the QD

surface are compensated by S2- ions, which enables facets to fuse via the newly formed

stoichiometric surfaces.

Pb2+(C17H33COO−)2 + Na2S −→ PbS + 2Na+C17H33COO− (3.2)

The S2- ion is classified as an X-type ligand according to the Covalent Bond Classification

(C.B.C.) developed by M.L.H. Green [12]. This type of ligand donates an electron to the

metal cation, and forms normal covalent bonds.

Despite the long carbon chain, oleates formed with monovalent cations like Na+ are water

soluble (100 mg/ml) and thus dissolve in the ethylene glycol subphase. The reaction

is controlled by the amount of available Na2S at the surface of the subphase, in turn


determined by the concentration of the reagent.

This route has the advantage that the superlattice is formed at room temperature, which

excludes convection current and allows the amount of material at the surface of the sub-

phase to be better controlled. This results in the formation of very homogeneous mono-

layers of the superlattice, extending over a large area.

Removal of the organic ligand shell was also achieved by using as the subphase either

ethylene glycol at 50 ◦C or diethylene glycol at room temperature. According to the

C.B.C. system, both glycols are classified as an L-type ligand. This type of ligand donates

two electrons to the metal cation in the form of a lone pair. In this way, a dative covalent

bond is formed. The difference in ligating strength between EG and DEG follows from

their structure. EG contains two OH groups with a total of four lone pairs. DEG has an

additional OR2 functionality, contributing two extra lone pairs.

Owen et al. showed that Pb(OA)2 units are adsorbed relatively weakly to the nanocrystal

surface and that L-type ligands induce removal of the metal carboxylates by coordinating

with the metal cation [25]. The experiments here indicate that this occurs with EG and

DEG. After the metal carboxylates are removed, the facets obtain stoichiometric surfaces

which enable interparticle bonding.

3.4.2 Oriented attachment

Small PbSe nanocrystals consist of 6 (100) facets and 12 (110) facets. Both the (100)

and (110) facets are evenly occupied by Pb and Se atoms. The (111) planes, however, are

either completely occupied by Pb or Se atoms. The difference in electronegativity of both

elements sets up a charge interaction in the crystal, which leads to a dipole moment. Both

the magnitude and the direction of the dipole moment is determined by the distribution

of Pb- and Se-occupied (111) planes. Cho et al. calculated for PbSe nanocrystals that a

dipole moment along a (100) axis not only has the highest probability, but also the highest

magnitude [4]. This leads to a preferential orientation and attachment of the nanocrystals

along this axis.


This is confirmed by the analysis of high resolution TEM images in this chapter. It was

shown that the majority of interparticle connections occur via (100) facets, and that bond-

ing via (110) only rarely occurs. Bonding via (111) facets was not observed at all.

Figure 3.24 compares the superlattice obtained by Evers et al. (A) [10] with the one obtained

in this thesis (B). In both superlattices the nanocrystals bond via (100) facets, although

the symmetry of the superlattice is clearly different.

For nanocrystals fusing via (100) planes, the bonding occurs between a Pb atom of one

crystal and a Se atom of the other. This means that the crystals are shifted at least half

a lattice parameter relative to each other. As a consequence, the angle between the (100)

bonding directions is not exactly 90◦, as would be the case for a perfect cubic symmetry.

In the superlattice obtained by Evers, bonding directions around 87◦and 81◦are seen, re-

spectively corresponding to lattice jumps of 0.5 and 1.5 times the lattice parameter. In the

superlattice obtained in this thesis, the measured angles are around 75◦, corresponding to

a shift of 2.5 times the lattice parameter. A factor that possibly contributes to this relative

large shift is the crystal shape of the QDs. If the nanocrystals are slightly elongated along

one (110) axis, opposite (100) facets are already intrinsically shifted relative to each other.

This means that even if the bonding occurs via a minimal shift of half a lattice parameter,

the overall symmetry of the superlattice corresponds to a seemingly larger shift. More

specifically, opposite (100) facets are shifted by 2 whole lattice parameters if a crystal is 24

% elongated along one (110) axis, starting from a perfectly symmetrical 6 nm nanocrystal.

The drawing in Figure 3.25 more visually illustrates the effect of elongation on the lattice

symmetry.


Figure 3.24: A: Cubic PbSe superlattice as obtained by Evers et al. B: PbSe superlattice

obtained in this thesis. Scale bars are 50 nm.

Figure 3.25: Schematic representation of a superlattice formed with perfect symmetrical

nanocrystals (left) and a superlattice formed with nanocrystals elongated along

a (110) axis. An elongation of 24 % corresponds to a shift of opposite (100) planes

by 2 lattice parameters.

3.4.3 Langmuir films

The Langmuir technique allows well defined monolayers to be fabricated over a large area,

with the extra advantage that the results are easily reproduced. When there is no in-

teraction between subphase and QDs, the controlled application of pressure enables the


formation of monolayers with the highest possible density. However, when the subphase is

adjusted in such a way that it removes ligands from the QD surface, this removal is induced

from the moment the nanocrystals are added to the trough. This leads to QDs partially

stripped from their ligands spread over the surface of the subphase. Upon compression,

nanocrystals in close proximity to each other bind together, leading to many small, rigid

groups of connected QDs, which in turn connect together when more pressure is applied.

Since the bonds are irreversible, faults in the stacking order cannot be corrected by rear-

rangement of individual nanocrystals, and this mechanism subsequently does not yield the

most dense stacking of QDs. It can be seen that delaying the ligand removal to a moment

where all nanocrystals are already in close proximity to each other, would lead to a more

dense superlattice. This was partly achieved by adding a small amount of DEG to the

subphase and letting the layer contract after a first initial compression. This improved

the quality of the layer and resulted in a homogeneous superlattice over a much larger

area, although the density of QDs could still be higher. More experiments are needed to

see whether it is possible to further increase the surface density by controlling the rate of

ligand removal.

Another possible approach is that in a first step, a hexagonal close packed layer is formed,

and in a second step, a reagent or an impulse is given to initiate ligand removal. It is for

example possible to heat up the subphase by recirculating hot water around the trough.

Alternatively, a reagent could be added from beneath the layer by homogeneous injection

into the subphase, or from above by the flow of a reactive gas. A bit more far fetched could

be the design of ligands which, for example, dissociate under influence of UV light.

3.4.4 Absorption enhancement

Although the number of absorbance spectra taken was insufficient to make conclusive

statements, the spectra shown in Figure 3.23 illustrate that the formation of interparticle

connections over a long range possibly enhances the absorption cross section of the QDs.

This is indicated by the increased cross section of connected QDs compared to unconnected


QDs (Table 3.1). The cross section was estimated by neglecting the reflection factor in

Equation 3.1. Although the correction for reflection is relatively small (< 10 %) for un-

connected QDs [11], no reflection data is available for connected QDs. This could lead to an

overall smaller cross section then the one reported here, but the numbers are only given as

an indication.

The increase of intrinsic absorption can be qualitatively understood from Equation 1.6.

For separate QDs, the intrinsic absorption coefficient is decreased by a partial screening of

the incident electrical field, determined by the local field factor. For bulk material, this is

not the case, and the local field factor equals 1. It can be rationalized that a connected

QD superlattice behaves as something between separate QDs and bulk material, and that

the local field factor is significantly increased compared to individual QDs.

3.5 Conclusion

This chapter shows that connected QD superlattices can be obtained via different tech-

niques.

Formation of superlattices at room temperature provides a good control over layer thick-

ness, and homogeneous monolayers over a large area were possible. By adding Na2S to the

subphase, the native oleate ligands were removed via an ion exchange reaction, and this

subsequently resulted in relatively dense superlattices of connected QDs.

Connected superlattices were also obtained at elevated temperature. Control over layer

thickness and homogeneity was more difficult, resulting in overall bilayered structures with

more cracks. The rate at which the nanocrystals were added to the subphase proved to be

a determining factor for the microstructure of the superlattice. The postulated mechanism

for ligand removal involves the loss of Pb(OA)2 units, leaving the QD surface largely

unpassivated.

Superlattices formed via Langmuir-Schaefer deposition on different subphases resulted in

crack-free films of varying density. For large areas, this technique provides a better control

of the homogeneity of the superlattice, which becomes important for reproducible device


fabrication. The experiments further showed that adjustment of the subphase composition

provides a way to influence and control the surface chemistry of the QDs. Two examples

were shown, where the native ligands were removed by either adding Na2S or DEG to the

subphase.

Finally, the absorption measurements indicated that the formation of a connected QD

superlattice possibly enhances the intrinsic absorption coefficient by increasing the local

field factor. This could be an extra advantage of using electronically coupled QDs for

photodetection.

Chapter 4

Quantum Dot Photodetectors

Several QD superlattices described in the previous chapter were tested for their photocon-

ductive properties by recording I-V curves of the superlattices deposited on gold electrodes.

By sweeping a voltage across the electrodes, a certain current is induced, depending on the

microstructure of the superlattice. This current was measured under different illumination

levels, ranging from 0 to 17684 W/m2, with a laser operating at a fixed wavelength of

1550 nm. From the resulting photocurrents, i.e. measured current minus dark current,

responsivities (A/W, calculated as R = Iphoto/Pincident) and contrast ratios (Iphoto/Idark)

were plotted in order to get a view on the performance of the detectors. This chapter only

shows the results from I-V measurements in the -5V to +5V range. I-V curves were also

recorded in the -10V to +10V range, although no significant changes were observed, apart

from an approximate doubling of the dark- and photocurrents.

Furthermore, the superlattices were subjected to a laser with a wavelength varying from

2100 nm to 2400 nm to see if photoconductivity was observed at energies below the band

gap of the individual QDs.

As a reference to the results further shown in this chapter, a superlattice consisting of a

hexagonal close packed QD layer with an intact organic ligand shell, as in Figure 3.16A

and B, shows no dark- or photocurrent, pointing to the effective insulating properties of

the organic shell. All other manipulations carried out during the superlattice formation

49

Chapter 4. Quantum Dot Photodetectors 50

proved successful and resulted in an increase of the conductivity.

4.1 Addition of Na2S to the subphase

Figure 4.1 shows the measured I-V curves at different illumination levels of the superlattices

formed with a 50 fold (A) and a 100 fold (B) excess Na2S in the subphase. Figure 4.2 shows

the corresponding photocurrents (A), responsivities (B) and contrast ratios Iphoto/Idark (C)

at 5 V in function of the illumination level.

As explained earlier, additional Na2S led to the removal of native oleate ligands, and a

significant increase of interparticle connections, related to the excess amount that was

added. This directly enhances the conducting behavior of the superlattice by reducing the

time a charge carrier needs to travel between the electrodes (τtransit). This is illustrated

by an increase of the dark current (at 5 V) from 0 µA to 10.5 µA going from no Na2S to

a 50 fold excess, and an increase from 10.5 µA to 29.0 µA going from a 50 fold to a 100

fold excess. This relative high dark current most likely points to a considerable doping of

the nanocrystals by the S2- ions.

The reduced transit time allows the superlattice to produce more photocurrent per incident

photon power. This results in a significant increase of both photocurrent and responsivity

by the addition of more Na2S. The maximum photocurrent increases from 3.2 µA to 10.0

µA at 800 µW for a 50 fold and a 100 fold excess Na2S respectively. The maximum

responsivity is found at 50 µW, where an increase from 0.017 A/W to 0.062 A/W is seen.

With the surface of the laser beam (240 µm diameter) and the density of particles, it is

possible to express a responsivity per QD, resulting in 6 · 10-11 A/W per QD for the 100 fold

excess superlattice. Figure 4.2C shows that both superlattices have a comparable contrast

ratio, suggesting that the observed increase in photocurrent and responsivity is mainly due

to the reduced transit time and not by an additional increase of the carrier lifetime.

No photoconductivity was observed between 2100 nm to 2400 nm.


Figure 4.1: Measured I-V curves of the superlattice formed with a 50 fold (A) and a 100 fold

(B) excess Na2S in the subphase.

Figure 4.2: Photocurrent (A), responsivity (B) and contrast ratio (C) of superlattices formed

with a 50 fold (red) and a 100 fold (blue) excess Na2S in the subphase.

4.2 Temperature of the subphase

Fast addition of QDs

Superlattice formation at elevated temperature, obtained by a fast addition of the QD

solution, resulted in a connected assembly of QDs (see Figure 3.5B). The measured I-V

curves are shown in Figure 4.3. Figure 4.4 shows the corresponding photocurrents (A),

responsivities (B) and contrast ratios Iphoto/Idark (C) at 5 V.


The superlattice shows a remarkable increase of conductivity, with a dark current (at 5 V)

of 209.8 µA. On one hand, this is due to a structure with a high density of interconnected

particles, thus bringing in more freely moving charge carriers. On the other hand, the

bilayered structure can serve as a bridge between cracks or isolated parts, thereby reducing

the percolation path and activating ’dead’ areas, which are otherwise effectively lost in a

monolayer. Additionally, doping by surface defects could further increase the dark current

by increasing the carrier density. Of all experiments described in this thesis, illumination

of this QD superlattice leads to the highest photocurrents, with a maximal generated

photocurrent of 63.0 µA at an incident optical power of 800 µW. Correspondingly, the

responsivity of this superlattice is also the highest one obtained, and a maximal responsivity

of 0.26 is found at 50 µW. Expressed per QD this becomes 13 · 10-11 A/W per QD. Figure

4.4C shows a maximal contrast ratio Iphoto/Idark around 0.3.

Going to longer wavelengths, this superlattice still shows a small photocurrent between

2100 nm and 2400 nm. The responsivities in this spectral region are shown in Figure 4.5.

Even at 2400 nm, there is still a little photocurrent, indicating that the absorption onset

starts at even longer wavelengths.

Figure 4.3: Measured I-V curves of the superlattice formed by fast addition of the QDs to the

subphase at 50 ◦C.



by fast addition of the QDs to the subphase at 50 ◦C.

Figure 4.5: Responsivity of the superlattice formed by fast addition of the QDs at 50 ◦C in the

2100 nm to 2400 nm spectral region, i.e. below the band gap energy of the colloidal

QDs.

Slow addition of QDs

Superlattice formation at elevated temperature, obtained by a drop-wise addition of the

QD solution, resulted in a molten-like structure (see Figure 3.11B). The measured I-V

curves are shown in Figure 4.6. Figure 4.7 shows the corresponding photocurrents (A),

responsivities (B) and contrast ratios Iphoto/Idark (C) at 5 V.


The structure shows a dark current (at 5V) of 35.9 µA. Although the electrical connections

in this structure are good, both the lower density of material and the much more hindered

percolation path of charge carriers contribute to a lower dark current. The structure

generates a photocurrent of 35.6 µA at 800 µW, with a corresponding contrast ratio of

0.99. A maximal responsivity of 0.29 A/W is found at 25 µW.

This structure showed a photoconductive response up untill 2300 nm. At 2400 nm, no

photocurrent was observed. The responsivities in this region are shown in Figure 4.8.

Figure 4.6: Measured I-V curves of the superlattice formed by drop-wise addition of the QDs

to the subphase at 50 ◦C.


by drop-wise addition of the QDs to the subphase at 50 ◦C.


Figure 4.8: Responsivity of the molten-like structure in the 2100 nm to 2400 nm spectral region,

i.e. below the band gap energy of the colloidal QDs.

4.3 Superlattices by Langmuir-Schaefer deposition

As stated earlier, the superlattices formed on pure water and ethylene glycol did not

show any conductivity. On the other hand, superlattices formed on a mixture of ethylene

glycol and diethylene glycol (Figure 3.18) and on a subphase containing Na2S (Figure

3.21) showed promising features and were homogeneous over a large area. Photodetectors

of these superlattices were made by depositing 4 layers via Langmuir-Schaefer deposition.

The measured I-V curves are shown in Figure 4.9. Figure 4.10 shows the corresponding

photocurrents (A), responsivities (B) and contrast ratios Iphoto/Idark (C) at 5 V.

Both superlattices have a very low dark conductivity, with dark currents (at 5 V) of 0.69

µA on the EG/DEG subphase and 0.79 µA on the Na2S containing subphase. Upon

illumination, photocurrents of several microamperes are generated, with maximal values

at 800 mW of 4.8 µA on the EG/DEG subphase and 7.5 µA on the Na2S containing

subphase. Correspondingly, the obtained responsivities in these superlattices are relatively

low, with the largest values found at 50 mW; 0.037 A/W for the EG/DEG subphase and

0.056 A/W for the Na2S containing subphase. Although the values of the photocurrents


are not very high, the low dark current is responsible for fairly high contrast ratios, the

highest ones obtained. At an incident optical power of 800 mW, the ratios are 6.9 for the

EG/DEG subphase and 9.5 for the Na2S containing subphase.

The Langmuir-Schaefer films did not show photoconductivity between 2100 nm and 2400

nm.

Figure 4.9: Measured I-V curves of superlattices made by 4 Langmuir-Schaefer depositions,

formed on a subphase containing 70 % EG and 30 % DEG (A), and on an EG

subphase containing a 100 fold excess Na2S (B).

Figure 4.10: Photocurrent (A), responsivity (B) and contrast ratio (C) of superlattices made

by 4 Langmuir-Schaefer depositions, formed on a subphase containing 70 % EG

and 30 % DEG (red curves), and on an EG subphase containing a 100 fold excess

Na2S (blue curves).


4.4 Discussion

4.4.1 Responsivity per QD

The responsivity per QD allows some comparison between different superlattices. Com-

paring the superlattice formed by fast addition at 50 ◦C to the superlattice formed by a

100 fold excess Na2S, it is seen that this quantity is about twice as high. At first sight this

can be explained by the better conductivity of the first, but since the dark current is about

7 times higher, another factor plays in favor of the latter. If the same level of doping and

quantum efficiency is assumed in the QDs, the responsivity scales with τlifetime/τtransit,

meaning that the lifetime of the carrier is about 3.5 times higher in the layer formed on

Na2S. This indicates that the S2- ions possibly introduce effective electron traps in the band

structure. Unless the structure consists of a significant amount of barriers or dead areas,

it is also possible that the S2- ions introduce some hole traps near the valence band, which

could partly explain the lower conductivity. However, these results are only indicative, and

more precise experiments aimed at determining the electronic structure of the superlattice

should provide more clarity.

4.4.2 Fast versus slow addition

The completely different morphology and their corresponding I-V characteristics of the

layers formed at 50 ◦C, i.e. the dense connected assembly of QDs and the less dense

molten-like structure, could provide some insights in properties influencing the photocon-

ductive behavior. Apart from a smaller amount of charge carriers present in the molten

structure, the transit time of the carriers is expected to be much higher due to the more

hindered, random percolation path. This can qualitatively explain the strongly reduced

dark current. Despite this, the relative generated photocurrent is significantly higher in the

molten structure, as illustrated by the increased Iphoto/Idark ratio. This is also shown by the

fact that both structures have a comparable responsivity, although the molten structure

has a much lower density.


The mechanism of ligand removal is identical for both structures and occurs most likely by

removal or displacement of complete Pb(OA)2 units, as discussed earlier. This leaves the

surface of the QDs largely unpassivated, possibly creating many surface trap states and

thereby increasing the carrier lifetime and leading to additional doping. In this perspective,

the relative larger photocurrent in the molten structure can be rationalized by both a larger

surface area and a more complete removal of the organic ligands (see FTIR spectra in Figure

3.8 and 3.13). Additionally, the complex structure of the molten-like superlattice can

possibly further increase the carrier lifetime by a principle of harder-to-find recombination

centers.

Both these structures show photoconductivity from 2100 nm to 2300 nm, which indicates

that the absorption onset has red shifted by several hundreds of nanometers compared to

the colloidal QDs. This could result in a higher absorption coefficient at 1550 nm, possibly

explaining the higher responsivity at this wavelength. This is also a promising feature for

shifting the spectral response of QD based photodetectors towards longer wavelengths.

4.4.3 Langmuir-Schaefer films

Explaining the I-V curves of the Langmuir-Schaefer films is a bit more difficult. Both films

show a very low dark current. However, upon illumination, photocurrents with relatively

high Iphoto/Idark ratios are generated, meaning that a conductive path is available for the

charge carriers. This indicates that few intrinsic mobile charge carriers are present in

the structure. Since both Langmuir-Schaefer films still contain a large amount of native

ligands, it is well possible that the oleates provide a better passivation and that in the

other methods the many created surface states lead to an effective doping of the QDs,

thereby increasing the dark current.

Since both the dark currents and the amount of organic ligands are comparable for the

two Langmuir-Schaefer films, the difference in photocurrent can be seen as an effect of the

surface density and the nature of the introduced trap states. In the case of the EG/DEG

subphase, the ligand removal is postulated as the loss of complete Pb(OA)2 units, leaving


part of the surface unpassivated. Since the Langmuir films were prepared under ambient

conditions, these sites are prone to oxidation, and partial passivation of the surface by

O2- ions or impurities of the subphase can take place. In the case of the EG subphase

containing Na2S, the surface is mainly passivated by oleate and S2- ions, which was shown

to increase the lifetime of the carrier. This allows more photocurrent to be generated,

increasing the responsivity and contrast ratio of the sulfur passivated QDs.

4.5 Conclusion

The I-V measurements in this chapter show that all manipulation made to remove the

native ligand shell improved the conductivity of the superlattices by providing connected

current pathways. The measurements further indicate that the precise nature of the surface

chemistry strongly influences the performance of the photodetector. More specifically,

passivation of the QD surface and controlled introduction of trap states are key factors for

enhancing the device performance.

An improved passivation of the QD surface ensures that the dark current remains low.

This has the advantage of enabling high contrast ratios, thereby increasing the sensitivity

of the device.

Introduction of trap states further increases the performance by increasing the exciton

lifetime. However, an important requirement is that the trap states do not act as a dopant.

This again increases the dark current and counteracts the advantage of passivation.

This thesis has shown that it is possible to make connected QD superlattices and that these

structures show promising features as photosensitive material. However, a more thorough

characterization and optimization of the chemical, and thus electronic, properties is needed.

Computational methods could prove a useful tool for screening and selecting possible trap

states or passivating agents. The use of a transistor type of setup would allow a more

precise study of the effect of carrier mobility and carrier density on the conductivity.

Appendix A

QD Langmuir-Blodgett/Schaefer

films

QD Langmuir films are QD monolayers formed by Langmuir-Blodgett/Schaefer deposition.

The films are produced on a Langmuir trough filled with a polar liquid, typically water.

The liquid in the trough is called the subphase. Due to the hydrophobic QD ligand shell, a

solution of QDs dispersed in a non-polar, volatile solvent can be dropcast on the subphase

without dissolving in it. As the solvent evaporates, the QDs evenly spread across the

surface of the subphase. At this stage, the distance between the QDs is relatively large.

By gently compressing the QDs with one or two moving barriers, the distance is gradually

decreased until eventually a dense, close packed monolayer is obtained.

The process of layer formation is monitored by careful measurement of the surface pressure.

This is achieved by partially immersing a Wilhelmy plate in the subphase and attaching

it to a microbalance. The surface pressure is determined from the variation of the force

exerted on the Wilhelmy plate via Equation A.1, where Π is the surface pressure and wp

and tp respectively the width and thickness of the Wilhelmy plate.

Π = − ∆F

2(wp + tp)(A.1)

The surface pressure is plotted in function of the available surface area and the resulting

pressure-area isotherm provides a way to monitor the formation of the close packed mono-

60

Appendix A. QD Langmuir-Blodgett/Schaefer films 61

layer. This is schematically shown in Figure A.1. At low pressure the QDs are regarded

as an ideal gas (G) due to the large interparticle distance. By gradually decreasing the

surface area, the interparticle distance decreases and the surface pressure increases. In this

process, the behavior of the QDs goes from gas-like (G) to liquid-like (L) to solid-like (S).

The formation of a close packed monolayer (S) is signaled by a very steep slope of the

isotherm.

Figure A.1: Pressure-area isotherm of the formation of a close packed monolayer. The behavior

of the QDs goes from gas-like (G) to liquid-like (L) to solid-like (S).

After compression, the film is transferred to a substrate of choice, either by vertically

pulling out a substrate, called Langmuir-Blodgett deposition, or by stamping the substrate

horizontally on the film, called Langmuir-Schaefer deposition (Figure A.2).

Figure A.2: Schematic representation of the Langmuir-Blodgett and Langmuir-Schaefer tech-

nique.

Bibliography

[1] Vesna Aleksandrovic, Denis Greshnykh, Igor Randjelovic, Andreas Fromsdorf, An-

dreas Kornowski, Stephan Volkher Roth, Christian Klinke, and Horst Weller. Prepa-

ration and electrical properties of cobalt-platinum nanoparticle monolayers deposited

by the langmuir-blodgett technique. ACS Nano, 2(6):1123–1130, 2008. ISSN 1936-

0851. doi: 10.1021/nn800147a. URL http://dx.doi.org/10.1021/nn800147a.

[2] William J. Baumgardner, Kevin Whitham, and Tobias Hanrath. Confined-but-

connected quantum solids via controlled ligand displacement. Nano Letters, 13(7):

3225–3231, 2013. ISSN 1530-6984. doi: 10.1021/nl401298s. URL http://dx.doi.

org/10.1021/nl401298s.

[3] M. Brust, R. Etchenique, E. J. Calvo, and G. J. Gordillo. The self-assembly of gold

and scd nanoparticle multilayer structures studied by quartz crystal microgravimetry.

Chemical Communications, (16):1949–1950, 1996. ISSN 1359-7345. doi: 10.1039/

CC9960001949. URL http://dx.doi.org/10.1039/CC9960001949.

[4] Kyung-Sang Cho, Dimitri V. Talapin, Wolfgang Gaschler, and Christopher B. Murray.

Designing pbse nanowires and nanorings through oriented attachment of nanoparticles.

J Am Chem Soc, 2005.

[5] J. J. Choi, C. R. Bealing, K. Bian, K. J. Hughes, W. Zhang, D. M. Smilgies, R. G.

Hennig, J. R. Engstrom, and T. Hanrath. Controlling nanocrystal superlattice sym-

metry and shape-anisotropic interactions through variable ligand surface coverage. J

62

http://dx.doi.org/10.1021/nn800147a

http://dx.doi.org/10.1021/nl401298s

http://dx.doi.org/10.1021/nl401298s

http://dx.doi.org/10.1039/CC9960001949

Bibliography 63

Am Chem Soc, 133(9):3131–8, 2011. ISSN 1520-5126 (Electronic) 0002-7863 (Linking).

doi: 10.1021/ja110454b. URL http://www.ncbi.nlm.nih.gov/pubmed/21306161.

[6] Angang Dong, Xingchen Ye, Jun Chen, Yijin Kang, Thomas Gordon, James M.

Kikkawa, and Christopher B. Murray. A generalized ligand-exchange strategy enabling

sequential surface functionalization of colloidal nanocrystals. Journal of the American

Chemical Society, 133(4):998–1006, 2010. ISSN 0002-7863. doi: 10.1021/ja108948z.

URL http://dx.doi.org/10.1021/ja108948z.

[7] Angang Dong, Yucong Jiao, and Delia J. Milliron. Electronically coupled

nanocrystal superlattice films by in situ ligand exchange at the liquid-air in-

terface. ACS Nano, 7(12):10978–10984, 2013. ISSN 1936-0851. doi: 10.

1021/nn404566b. URL http://pubs.acs.org/doi/abs/10.1021/nn404566bhttp:

//pubs.acs.org/doi/pdfplus/10.1021/nn404566b.

[8] M. D. Eldridge, P. A. Madden, and D. Frenkel. Entropy-driven formation of a su-

perlattice in a hard-sphere binary mixture. Nature, 365(6441):35–37, 1993. URL

http://dx.doi.org/10.1038/365035a0.

[9] M. D. Eldridge, P. A. Madden, and D. Frenkel. The stability of the ab13 crys-

tal in a binary hard sphere system. Molecular Physics, 79(1):105–120, 1993. ISSN

0026-8976. doi: 10.1080/00268979300101101. URL http://dx.doi.org/10.1080/

00268979300101101.

[10] W. H. Evers, B. Goris, S. Bals, M. Casavola, J. de Graaf, R. van Roij, M. Dijkstra, and

D. Vanmaekelbergh. Low-dimensional semiconductor superlattices formed by geomet-

ric control over nanocrystal attachment. Nano Letters, 13(6):2317–2323, 2013. ISSN

1530-6984. doi: 10.1021/nl303322k. URL <GotoISI>://WOS:000320485100001http:

//pubs.acs.org/doi/abs/10.1021/nl303322k.

[11] Pieter Geiregat, Yolanda Justo, Sofie Abe, Stijn Flamee, and Zeger Hens. Giant and

broad-band absorption enhancement in colloidal quantum dot monolayers through

http://www.ncbi.nlm.nih.gov/pubmed/21306161

http://dx.doi.org/10.1021/ja108948z

http://pubs.acs.org/doi/abs/10.1021/nn404566b http://pubs.acs.org/doi/pdfplus/10.1021/nn404566b

http://pubs.acs.org/doi/abs/10.1021/nn404566b http://pubs.acs.org/doi/pdfplus/10.1021/nn404566b

http://dx.doi.org/10.1038/365035a0

http://dx.doi.org/10.1080/00268979300101101

http://dx.doi.org/10.1080/00268979300101101

<Go to ISI>://WOS:000320485100001 http://pubs.acs.org/doi/abs/10.1021/nl303322k

<Go to ISI>://WOS:000320485100001 http://pubs.acs.org/doi/abs/10.1021/nl303322k

Bibliography 64

dipolar coupling. ACS Nano, 7(2):987–993, 2013. ISSN 1936-0851. doi: 10.1021/

nn305524a. URL http://dx.doi.org/10.1021/nn305524ahttp://pubs.acs.org/

doi/pdfplus/10.1021/nn305524a.

[12] M. L. H. Green. A new approach to the formal classification of covalent compounds

of the elements. Journal of Organometallic Chemistry, 500(1-2):127–148, 1995. ISSN

0022-328X. doi: http://dx.doi.org/10.1016/0022-328X(95)00508-N.

[13] Tobias Hanrath, Dirk Veldman, Joshua J. Choi, Christina G. Christova, Martijn M.

Wienk, and Rene A. J. Janssen. Pbse nanocrystal network formation during pyridine

ligand displacement. ACS Applied Materials & Interfaces, 1(2):244–250, 2009. ISSN

1944-8244. doi: 10.1021/am8001583. URL http://dx.doi.org/10.1021/am8001583.

[14] Zeger Hens and Jose C. Martins. A solution nmr toolbox for characterizing the sur-

face chemistry of colloidal nanocrystals. Chemistry of Materials, 25(8):1211–1221,

2013. doi: 10.1021/cm303361s. URL http://pubs.acs.org/doi/abs/10.1021/

cm303361s.

[15] Zeger Hens and Iwan Moreels. Light absorption by colloidal semiconductor quantum

dots. J. Mater. Chem., 22:10406–10415, 2012. doi: 10.1039/C2JM30760J. URL

http://dx.doi.org/10.1039/C2JM30760J.

[16] Chen Hu, Alban Gassenq, Yolanda Justo, Sergii Yakunin, Wolfgang Heiss, Zeger Hens,

and Gunther Roelkens. Short-wave infrared colloidal quantum dot photodetectors on

silicon. volume 8631, pages 863127–863127–8. URL http://dx.doi.org/10.1117/

12.2001246.

[17] M. V. Jarosz, V. J. Porter, B. R. Fisher, M. A. Kastner, and M. G. Bawendi. Pho-

toconductivity studies of treated cdse quantum dot films exhibiting increased ex-

citon ionization efficiency. Physical Review B, 70(19):195327, 2004. URL http:

//link.aps.org/doi/10.1103/PhysRevB.70.195327.

http://dx.doi.org/10.1021/nn305524a http://pubs.acs.org/doi/pdfplus/10.1021/nn305524a

http://dx.doi.org/10.1021/nn305524a http://pubs.acs.org/doi/pdfplus/10.1021/nn305524a

http://dx.doi.org/10.1021/am8001583

http://pubs.acs.org/doi/abs/10.1021/cm303361s

http://pubs.acs.org/doi/abs/10.1021/cm303361s

http://dx.doi.org/10.1039/C2JM30760J

http://dx.doi.org/10.1117/12.2001246

http://dx.doi.org/10.1117/12.2001246

http://link.aps.org/doi/10.1103/PhysRevB.70.195327

http://link.aps.org/doi/10.1103/PhysRevB.70.195327

Bibliography 65

[18] Y. Justo, I. Moreels, K. Lambert, and Z. Hens. Langmuir-blodgett monolayers

of colloidal lead chalcogenide quantum dots: morphology and photoluminescence.

Nanotechnology, 21(29):295606, 2010. ISSN 1361-6528 (Electronic) 0957-4484 (Link-

ing). doi: 10.1088/0957-4484/21/29/295606. URL http://www.ncbi.nlm.nih.gov/

pubmed/20601759.

[19] C. J. Kiely, J. Fink, M. Brust, D. Bethell, and D. J. Schiffrin. Spontaneous ordering

of bimodal ensembles of nanoscopic gold clusters. Nature, 396(6710):444–446, 1998.

ISSN 0028-0836. URL http://dx.doi.org/10.1038/24808.

[20] Karel Lambert, Richard K. Capek, Maryna I. Bodnarchuk, Maksym V. Kovalenko,

Dries Van Thourhout, Wolfgang Heiss, and Zeger Hens. Langmuir-schaefer deposi-

tion of quantum dot multilayers. Langmuir, 26(11):7732–7736, 2010. ISSN 0743-

7463. doi: 10.1021/la904474h. URL http://dx.doi.org/10.1021/la904474hhttp:

//pubs.acs.org/doi/abs/10.1021/la904474h.

[21] Iwan Moreels, Karel Lambert, David De Muynck, Frank Vanhaecke, Dirk Poelman,

Jose C. Martins, Guy Allan, and Zeger Hens. Composition and size-dependent ex-

tinction coefficient of colloidal pbse quantum dots. Chemistry of Materials, 19(25):

6101–6106, 2007. doi: 10.1021/cm071410q. URL http://pubs.acs.org/doi/abs/

10.1021/cm071410q.

[22] C. B. Murray, C. R. Kagan, and M. G. Bawendi. Self-organization of

cdse nanocrystallites into three-dimensional quantum dot superlattices.

Science, 270(5240):1335–1338, 1995. doi: 10.1126/science.270.5240.1335.

URL http://www.sciencemag.org/content/270/5240/1335.abstracthttp:

//www.sciencemag.org/content/270/5240/1335.

[23] A. Nag, M. V. Kovalenko, J. S. Lee, W. Liu, B. Spokoyny, and D. V. Talapin. Metal-

free inorganic ligands for colloidal nanocrystals: S2-, hs-, se2-, hse-, te2-, hte-, tes3(2-

), oh-, and nh2- as surface ligands. J Am Chem Soc, 133(27):10612–20, 2011. ISSN



http://dx.doi.org/10.1038/24808

http://dx.doi.org/10.1021/la904474h http://pubs.acs.org/doi/abs/10.1021/la904474h

http://dx.doi.org/10.1021/la904474h http://pubs.acs.org/doi/abs/10.1021/la904474h

http://pubs.acs.org/doi/abs/10.1021/cm071410q

http://pubs.acs.org/doi/abs/10.1021/cm071410q

http://www.sciencemag.org/content/270/5240/1335.abstract http://www.sciencemag.org/content/270/5240/1335

http://www.sciencemag.org/content/270/5240/1335.abstract http://www.sciencemag.org/content/270/5240/1335

Bibliography 66

1520-5126 (Electronic) 0002-7863 (Linking). doi: 10.1021/ja2029415. URL http:

//www.ncbi.nlm.nih.gov/pubmed/21682249.

[24] Soong Ju Oh, Nathaniel E. Berry, Ji-Hyuk Choi, E. Ashley Gaulding, Hangfei

Lin, Taejong Paik, Benjamin T. Diroll, Shin Muramoto, Christopher B. Mur-

ray, and Cherie R. Kagan. Designing high-performance pbs and pbse nanocrys-

tal electronic devices through stepwise, post-synthesis, colloidal atomic layer de-

position. Nano Letters, 14(3):1559–1566, 2014. ISSN 1530-6984. doi: 10.1021/

nl404818z. URL http://dx.doi.org/10.1021/nl404818zhttp://pubs.acs.org/

doi/pdfplus/10.1021/nl404818z.

[25] Jonathan S. Owen, Jungwon Park, Paul-Emile Trudeau, and A. Paul Alivisatos.

Reaction chemistry and ligand exchange at cadmium-selenide nanocrystal surfaces.

Journal of the American Chemical Society, 130(37):12279–12281, 2008. ISSN 0002-

7863. doi: 10.1021/ja804414f. URL http://dx.doi.org/10.1021/ja804414fhttp:

//pubs.acs.org/doi/pdfplus/10.1021/ja804414f.

[26] F. X. Redl, K. S. Cho, C. B. Murray, and S. O’Brien. Three-dimensional binary

superlattices of magnetic nanocrystals and semiconductor quantum dots. Nature,

423(6943):968–971, 2003. ISSN 0028-0836. doi: http://www.nature.com/nature/

journal/v423/n6943/suppinfo/nature01702 S1.html. URL http://dx.doi.org/10.

1038/nature01702.

[27] E. L. Rosen, R. Buonsanti, A. Llordes, A. M. Sawvel, D. J. Milliron, and B. A. Helms.

Exceptionally mild reactive stripping of native ligands from nanocrystal surfaces by

using meerwein’s salt. Angew Chem Int Ed Engl, 51(3):684–9, 2012. ISSN 1521-

3773 (Electronic) 1433-7851 (Linking). doi: 10.1002/anie.201105996. URL http:

//www.ncbi.nlm.nih.gov/pubmed/22147424.

[28] E. V. Shevchenko, D. V. Talapin, S. O’Brien, and C. B. Murray. Polymorphism in

ab(13) nanoparticle superlattices: an example of semiconductor-metal metamaterials.



http://dx.doi.org/10.1021/nl404818z http://pubs.acs.org/doi/pdfplus/10.1021/nl404818z

http://dx.doi.org/10.1021/nl404818z http://pubs.acs.org/doi/pdfplus/10.1021/nl404818z

http://dx.doi.org/10.1021/ja804414f http://pubs.acs.org/doi/pdfplus/10.1021/ja804414f

http://dx.doi.org/10.1021/ja804414f http://pubs.acs.org/doi/pdfplus/10.1021/ja804414f

http://dx.doi.org/10.1038/nature01702

http://dx.doi.org/10.1038/nature01702



Bibliography 67

J Am Chem Soc, 127(24):8741–7, 2005. ISSN 0002-7863 (Print) 0002-7863. doi:

10.1021/ja050510z.

[29] Paul Simon, Lydia Bahrig, Igor A. Baburin, Petr Formanek, Falk Roder, Jan Sick-

mann, Stephen G. Hickey, Alexander Eychmuller, Hannes Lichte, Rudiger Kniep, and

Elena Rosseeva. Interconnection of nanoparticles within 2d superlattices of pbs/oleic

acid thin films. Advanced Materials, pages n/a–n/a, 2014. ISSN 1521-4095. doi:

10.1002/adma.201305667. URL http://dx.doi.org/10.1002/adma.201305667.

[30] Ajay Singh, Robert D. Gunning, Shafaat Ahmed, Christopher A. Barrett, Niall J.

English, Jose-Antonio Garate, and Kevin M. Ryan. Controlled semiconductor

nanorod assembly from solution: influence of concentration, charge and solvent na-

ture. Journal of Materials Chemistry, 22(4):1562–1569, 2012. ISSN 0959-9428.

doi: 10.1039/C1JM14382D. URL http://dx.doi.org/10.1039/C1JM14382Dhttp:

//pubs.rsc.org/en/Content/ArticleLanding/2012/JM/c1jm14382d.

[31] Dmitri V. Talapin and Christopher B. Murray. Pbse nanocrystal solids for n- and

p-channel thin film field-effect transistors. Science, 310(5745):86–89, 2005. doi: 10.

1126/science.1116703. URL http://www.sciencemag.org/content/310/5745/86.

abstracthttp://www.sciencemag.org/content/310/5745/86.full.pdf.

[32] T. Wen and S. A. Majetich. Ultra-large-area self-assembled monolayers of nanopar-

ticles. ACS Nano, 5(11):8868–76, 2011. ISSN 1936-086X (Electronic) 1936-0851

(Linking). doi: 10.1021/nn2037048. URL http://www.ncbi.nlm.nih.gov/pubmed/

22010827.

[33] Andrew W. Wills, Moon Sung Kang, Ankur Khare, Wayne L. Gladfelter, and David J.

Norris. Thermally degradable ligands for nanocrystals. ACS Nano, 4(8):4523–4530,

2010. ISSN 1936-0851. doi: 10.1021/nn100637u. URL http://dx.doi.org/10.1021/

nn100637u.

http://dx.doi.org/10.1002/adma.201305667

http://dx.doi.org/10.1039/C1JM14382D http://pubs.rsc.org/en/Content/ArticleLanding/2012/JM/c1jm14382d

http://dx.doi.org/10.1039/C1JM14382D http://pubs.rsc.org/en/Content/ArticleLanding/2012/JM/c1jm14382d

http://www.sciencemag.org/content/310/5745/86.abstract http://www.sciencemag.org/content/310/5745/86.full.pdf

http://www.sciencemag.org/content/310/5745/86.abstract http://www.sciencemag.org/content/310/5745/86.full.pdf



http://dx.doi.org/10.1021/nn100637u

http://dx.doi.org/10.1021/nn100637u

Two -dimensional superlattices of colloidal quantum dots ...lib.ugent.be/fulltxt/RUG01/002/163/541/RUG01-002163541_2014_0001_… · PbSe quantum dot synthesis PbSe semiconductor nanocrystals

Documents