Ultrahigh-gain colloidal quantum dot infrared photodetectors: Unraveling the potential of electro-kinetically pumped charge multiplication

doi:10.21203/rs.3.rs-4109110/v1

Download PDF

Research Article

Ultrahigh-gain colloidal quantum dot infrared photodetectors: Unraveling the potential of electro-kinetically pumped charge multiplication

https://doi.org/10.21203/rs.3.rs-4109110/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Colloidal quantum dots (CQDs) are promising candidates for infrared photodetectors (IRPDs) with high detectivity (D^*) and low-cost production. However, the incoherent hopping of charge carriers often causes low carrier mobility and inefficient charge extraction, leading to low detectivity in CQD-based IRPDs. Although photo-induced charge multiplication, in which high-energy photons create multiple electrons, is a viable alternative for enhancing the signal amplitude and detectivity, its capability is limited in IR detectors because of its susceptibility to thermal noise in low-bandgap materials. Herein, we present, for the first time, a pioneering architecture of a CQD-based IRPD that employs kinetically pumped charge multiplication. This is achieved by employing a thick CQD layer (> 540 nm) and subjecting it to a strong electric field. This configuration accelerates electrons to acquire kinetic energy, surpassing the bandgap of the CQD material, thereby initiating kinetically pumped charge multiplication. We also demonstrate that optimizing the dot-to-dot distance to approximately 4.1 nm yields superior device performance because of the tradeoff between increased impact ionization rates and diminished electron-hopping probabilities with increasing dot-to-dot distance. The optimal CQD-based IRPD exhibited a maximum multiplication gain of 85 and a peak detectivity (D^*) of 1.4×10¹⁴ Jones at a wavelength of 940 nm.

Colloidal quantum dot

infrared photodetector

charge multiplication

electron hopping

impact ionization

dot-to-dot distance

With the advent of the Fourth Industrial Revolution, seamless interconnection and communication between humans and machines have been emphasized. In this regard, highly sensitive infrared (IR) detectors have become essential for bridging human–machine interfaces^1–6. Conventional IR photodetectors (IRPDs) containing silicon⁷, germanium⁸, and III–V compound materials⁹ require precise composition engineering for high performance. Therefore, they inevitably require a high-temperature epitaxy process, which increases the manufacturing complexity and cost^10,11.

Colloidal quantum dots (CQDs) have emerged as practical candidates for IRPDs owing to their solution processability and facile bandgap tunability in the IR region^12,13. However, typical CQD-based IRPDs have 10⁴–10⁶-fold lower carrier mobilities than bulk crystalline semiconductors because incoherent hopping of the charge carriers is dominant in CQD-based IRPDs^14,15. Moreover, the presence of dangling bonds on a CQD surface leads to increased defect density, which in turn causes charge recombination and hinders efficient charge extraction¹⁶. Consequently, thinner CQD films tend to improve the electrical properties of CQD-based IRPDs, despite the reduced IR absorption.

Various strategies, including the preparation of highly mono-dispersed CQDs¹⁷, proposals for a stable CQD ink formulation process for perfect surface passivation¹⁸, and applications of novel ligands for increased carrier mobility, have been put forward to reduce the noise current in CQD-based IRPDs¹⁹. Recently, halide- and thiol-based ligands have dramatically improved the carrier mobility by reducing the dot-to-dot distance (d_DtoD), thus increasing the hopping probability in CQD photovoltaics and photodetectors^20,21. However, the Coulombic force intensifies the CQD interactions as d_DtoD decreases, leading to their aggregation and consequently reduced detectivity and stability.

Alternatively, signal amplification techniques based on impact ionization for electron multiplication have been proposed for conventional CQD. For instance, when photons impart energy to electrons exceeding the bandgap of the material, secondary electrons are formed by interacting with the excess potential energy and are subsequently transferred to adjacent CQDs^22,23. However, these phenomena may be inappropriate for IRPDs because low-bandgap CQDs are susceptible to thermal noise, which can compromise the detectivity in CQD-based IRPDs.

Herein, we propose a new strategy that uses kinetic energy to facilitate charge multiplication in high-detectivity CQD-based IRPDs. Accelerated by strong electric fields within the CQD layer, electrons gain kinetic energy and subsequently transfer this excess energy to neighboring valence electrons through electron–electron collisions, thereby generating multiple charge carriers within the CQD layer. In CQD-based IRPDs with a 585 nm-thick CQD layer, we observed a positive temperature-dependent breakdown voltage behavior, suggesting the occurrence of kinetically pumped charge multiplication. Our multi-physics models support the hypothesis that the primary mechanism of current generation shifts from electron tunneling to impact ionization in CQD-based IRPDs when the CQD layer thickness exceeds 540 nm. Further evidence from density functional theory (DFT) calculations and operando photoluminescence (PL) measurements revealed that an increase in d_DtoD, modulated by the ligand chain length, lowers the threshold energy required for charge multiplication. However, this increase in d_DtoD simultaneously impedes electron hopping in CQD-based IRPDs. Consequently, we engineered a charge-multipliable CQD-based IRPD that exhibits a maximum gain of 85, a specific detectivity of 1.4×10¹⁴ Jones, and a bandwidth of 1.1×10⁶ Hz at a wavelength of 940 nm, outperforming the emerging solution-processable IRPDs.

CQD layer thickness effect on charge multiplication

To investigate the prerequisites for initiating charge multiplication via electrokinetic pumping, we studied the effect of the layer thickness of the CQD. Typically, a sufficiently thick CQD layer is crucial for attaining the kinetic energy required to derive the impact ionization in CQD-based IRPDs^24–27. The device architecture included zinc oxide (ZnO) as the electron-transporting layer, with indium tin oxide (ITO) and gold (Au) serving as the cathode and anode, respectively. The overall structure of our CQD-based IRPD, as revealed by cross-sectional transmission electron microscopy (TEM), was ITO (75 nm)/ZnO (30 nm)/CQD (585 nm)/Au (50 nm) (Fig. 1a). Scanning TEM energy dispersive spectroscopy (STEM–EDS) revealed distinct interfaces, indicating the successful construction of CQD-based devices without chemical degradation (Figure S1).

The activation energy for the CQD-based IRPDs was determined from the temperature-dependent dark current at various reverse bias points using the Arrhenius model (Figure S2). The device exhibited a potential barrier of approximately 0.33 eV in our devices, which hinders the charge transfer in accordance with previous studies^28,29 (Fig. 1b). The potential barrier height decreased with the increase in the reverse bias in the CQD-based IRPDs. Upon reaching a reverse bias of 10 V, the activation energy transitioned from positive to negative, indicating that the internal potential assisted the transfer of conduction electrons in the CQD-based IRPDs³⁰. When the reverse bias exceeded 14 V, a noticeable increase in the current was observed, despite the saturation of the activation energy. This suggests that in the high-reverse-bias regions (V > 14 V), a notable amount of current is generated from charge multiplication rather than merely by efficient charge transfer.

We examined how the breakdown voltage varied with the device temperature to identify the origin of charge multiplication in CQD-based IRPDs with varying CQD thicknesses (Figure S3 and Fig. 1c)³¹. All the CQD-based IRPDs exhibited a significant increase in the current when operating in high-reverse-bias regions exceeding 10 V. Figure 1c shows that the CQD-based IRPDs with CQD thicknesses of 360 and 495 nm have a negative correlation between the breakdown voltage and temperature, suggesting that increased thermal energy facilitates the tunneling of valence electrons, leading to an increased current³². In contrast, the 585 nm-thick CQD-based IRPD exhibited a positive correlation (Fig. 1c). This implies that increased thermal energy contributes to the vibrational excitation of valence electrons within the CQD layer, which in turn hinders electron acceleration by inducing collisions with the conduction electrons³³. As a result, a higher reverse bias is necessary to induce charge multiplication when the temperature rises because electrons must gain kinetic energy exceeding the bandgap before colliding with the adjacent valence electrons within the CQD layer^9,34. This positive correlation confirms that the increase in the current, observed in Fig. 1b, originates from kinetically pumped charge multiplication rather than electron tunneling.

To gain a deeper understanding of the charge multiplication behavior in CQD-based IRPDs, we established a 2D numerical model using Silvaco Atlas (Fig. 1d, Figure S4-6, and Supplementary Note 1). The Poole–Frenkel mobility model and Selberherr’s impact ionization model were employed to identify the charge multiplication in CQD-based devices^35,36. The material parameters were derived from experimental measurements and previous literature (Figure S4-5 and Table S1). With a reverse bias exceeding 24 V, impact ionization was observed to occur across a wide region of the CQD layer (≈ 585 nm) (Figure S6a). The rate of impact ionization increases as the electrons approach the ZnO layer, implying that the electrons gain kinetic energy as they traverse the devices. An increase in the impact ionization rate was also observed with increasing reverse bias. This trend was particularly pronounced when the reverse bias reached 26.8 V, at which point a marked increase in the impact ionization rate was observed (dotted line, Figure S6a). Figure S6b shows that at reverse biases exceeding 28 V, electron tunneling can be observed at the CQD–ZnO interfaces, indicating that tunneling electrons can contribute to current generation under high-reverse-bias conditions.

Next, we calculated the contribution of impact ionization to the current generation by considering both the applied reverse bias and CQD thicknesses in the CQD-based IRPDs (Fig. 1d). The breakdown voltages for both impact ionization (V_{BR, A}) and tunneling (V_{BR, T}) were extracted from Figure S7, which shows the calculated current densities for electron tunneling at the ZnO–CQD interfaces and the kinetically pumped charge multiplication within the CQD layer in the CQD-based IRPD with varying CQD thicknesses. The region where neither electron tunneling nor charge multiplication occurred was identified and defined as the neutral region (gray area in Fig. 1d). In CQD-based IRPDs with a CQD thickness of 200 nm, electron tunneling initiated a voltage of 5.7 V prior to the onset of impact ionization at the ZnO–CQD interfaces. This led to a dominant current contribution (≥ 97%) from electron tunneling, while a limited amplification contribution (≤ 5%) was observed even when the reverse bias increased beyond V_{BR, A} (23.2 V) in the CQD-based IRPDs. Upon further applying a reverse bias exceeding 27.2 V, the amplification contribution surpassed 50%, indicating that impact ionization became the dominant factor in the current generation. However, an excessively high current increases the device temperature, causing degradation and potential loss of functionality in our CQD-based IRPDs under such high-bias conditions. Hence, this region was designated as the impractical region (red hatch area in Fig. 1d). We also observed a transition in the dominant mechanism of current generation in the thicker CQD layers. Upon reaching a CQD thickness of 540 nm in the CQD-based IRPDs, kinetically pumped charge multiplication became the primary contributor before the onset of electron tunneling (i.e., V_{BR, A} < V_{BR, T}). As the CQD thickness was further increased beyond 585 nm in CQD-based IRPDs, an amplification contribution of approximately 75% was observed at V_{BR, A}. When the applied reverse bias exceeded V_{BR, A} by 1.2 V (i.e., 30 V), the amplification contribution accounted for more than 90%. We further noted a minor contribution to the current from charge multiplication at voltages even below V_{BR, A}. However, this region was designated as the low-multiplication region (light gray area in Fig. 1d) because of its negligible current contribution, particularly when compared with the current observed at voltages exceeding V_{BR, A}.

Figures 1e–f schematically illustrate the effect of the CQD layer thickness on the current generation mechanisms. In the low-reverse-bias regions below V_BR, the current in the CQD-based IRPDs was generated solely through photon absorption. When V_BR was exceeded, a narrow energy band at the CQD/ZnO interface prompted the tunneling of the valence electrons within the CQD layer, leading to a substantial increase in the current in the CQD-based IRPDs with a thin CQD layer, specifically in those with thicknesses less than 540 nm (Fig. 1e). However, this tunneling current served as dark current noise because it was generated independently of photon absorption. In contrast, for the CQD-based IRPDs with a thick CQD layer exceeding 540 nm, the photogenerated electrons were kinetically accelerated as they traversed the CQD layer under a reverse bias above V_BR (Fig. 1f). When these free electrons gain kinetic energy, they collide with other valence electrons within the CQD layer, inducing a transition from the valence to conduction bands and, as a result, a substantial increase in the current. Given that this current is generated by the multiplication of photogenerated electrons, it can enhance the signal amplitude in CQD-based IRPDs.

Characterization of the distances between CQD nanoparticles

We examined CQD solids treated with four types of thiol ligands with different carbon chain lengths to demonstrate the relationship between the dot-to-dot distance (d_DtoD) and charge carrier dynamics. The thiol ligands used were 1,2-ethanedithiol (EthaneDT), 1,3-propanedithiol (PropaneDT), 1,5-pentanedithiol (PentaneDT), and 1,8-octanedithiol (OctaneDT). High-resolution transmission electron microscopy (HRTEM) was employed to identify the effects of thiol treatment on CQD aggregation. The lattice parameter of the CQD was consistently observed to be 3.0 Å, corresponding to the (200) plane of the CQD structure³⁷. This implies that all the CQD films were successfully exchanged from native organic ligands to thiol ligands without undesirable CQD fusion (Fig. 2a, left).

The nanocrystalline superlattice (SL) structure was characterized using grazed-incidence small-angle X-ray scattering (GISAXS) (Fig. 2b). The OctaneDT-treated CQD film exhibited a broad scattering spectrum, indicating the construction of a relatively amorphous stacking structure compared with the other thiol ligands. In the thiol-treated CQD solids, a reduction in the length of the ligands led to a more pronounced (200)_SL peak relative to the (111)_SL peak. This suggests that more compact structures were formed with the transformation from a body-centered cubic (BCC) structure to a face-centered cubic (FCC) structure^38,39.

The average d_DtoDs of the CQD films treated with different thiol ligands were extracted from the azimuthal integration of the diffraction pattern (Fig. 2c)⁴⁰. The CQD solids treated with EthaneDT exhibited the most compact arrangement, with a d_DtoD of 3.86 nm. A small increase (approximately 0.2 nm) in d_DtoD was observed in the CQD solids treated with PropaneDT with three carbon chains. When the carbon chain length of the thiol ligands was extended to 8 in the case of the OctaneDT-treated CQD solids, d_DtoD increased significantly to 5.48 nm.

X-ray photoelectron spectroscopy (XPS) measurements were conducted to analyze the elemental composition of the CQD films depending on the thiol ligands (Fig. 2d and Figure S8). The presence of a thiol peak (163.70 eV) could be observed in the spectra of all the CQD films, confirming the successful replacement with thiol ligands (Figure S8a). Single carbon bonding at 284.5 eV also became more pronounced in the CQD films with the increase in the carbon chain length of the thiol ligands (Fig. 2d, left). Moreover, surface treatments with longer thiol ligands increased Pb-O peaks (143.2 eV and 138.5 eV) (Fig. 2d right and Figure S8b). These peaks can be attributed to the oxidation of dangling bonds on the CQD surface⁴¹, suggesting an increase in the number of dangling bonds in the CQD solids treated with longer thiol ligands (Supplementary Note 2). Table S2 shows that longer thiol ligands tend to have poor coverage on the CQD surfaces, possibly because of the increased steric hindrance during surface treatment^42,43.

Ultraviolet photoelectron spectroscopy (UPS) was used to evaluate the energy levels of the thiol-treated CQD solids (Figure S4b and Fig. 2e). Across all the thiol-treated CQD films, consistent conduction and valence band positions were observed. Nonetheless, variations in the Fermi level (E_F) within the bandgap of CQD solids were found to depend on the ligand chain length. Specifically, the CQD solids treated with EthaneDT exhibited a p-type behavior, with the Fermi level located below the midpoint of the band gap. Elongation of the carbon chain of the thiol ligands led to a shift in the E_F position toward the conduction band. This E_F shift in the CQD films could be influenced by the reduced coverage with longer ligands⁴⁴. Moreover, it has been previously reported that the dipole moment at the CQD–ligand interfaces increased with the elongation of thiol ligand chains in CQD films^45,46, potentially facilitating electron creation, thereby inducing an E_F shift toward the conduction band⁴⁷.

Next, we revealed the charge transfer characteristics with varying d_DtoDs using a photoluminescence (PL) spectroscopy analysis of our complete device structures (Fig. 2f). A longer d_DtoDs leads to an increase in the PL spectra of the CQD-based IRPD. When d_DtoDs reached 5.48 nm (i.e., OctaneDT), a more than 225-fold increase in the PL intensity was observed compared with the EthaneDT-treated device. The observed PL quenching in CQD solids with shorter d_DtoDs suggests superior charge-transport capabilities within the CQD layer.

To further clarify the interplay between charge multiplication and d_DtoDs, we conducted operando PL measurements on our CQD-based IRPD under varying reverse biases with a 785 nm pump wavelength prior to the irreversible breakdown of the devices (Fig. 2g and Figure S9). The normalized PL intensity (I_norm) was determined by dividing the peak PL intensity at each reverse bias by the corresponding value at 0 V. Peak PL intensities were observed at 992, 968, 962, and 970 nm for the EthaneDT-, PropaneDT-, PentaneDT-, and OctaneDT-treated devices, respectively. I_norm exhibited an initial decline with increasing reverse bias in all the CQD-based IRPDs, indicating that the applied electric field facilitates charge extraction in the CQD layer. Moreover, a marked increase in I_norm was observed when the reverse bias exceeded specific critical thresholds: 17 V for EthaneDT, 19.5 V for PropaneDT, 23.5 V for PentaneDT, and 40 V for OctaneDT. We hypothesized that the increase in I_norm is attributable to the initiation of charge multiplication mechanisms coupled with inferior charge transport properties, culminating in elevated rates of radiative recombination under these conditions. Therefore, this increase appears to be more pronounced with increasing d_DtoDs within the CQD-based IRPD. For example, a notable increase in I_norm of 180% was recorded for the OctaneDT-treated device.

Theoretical calculations for CQD nanoparticles depending on dot-to-dot distances

We developed density functional theory (DFT) models based on the results obtained from the characterization of CQD solids with different d_DtoDs values to understand the charge transport and multiplication dynamics among CQD particles (Fig. 3a and Supplementary Note 3). For this purpose, we utilized CQDs with a diameter of 2.1 nm, as confirmed through TEM and UV–Vis measurements (Figs. 2a and 2e). All the thiol-treated CQD particles exhibited the density of state characteristics of semiconductors devoid of mid-gap states (Figure S10). The threshold energy (E_th) was calculated to elucidate the impact ionization process in each thiol-treated CQD (Fig. 3b and supplementary Note 4)⁴⁸. Upon photon absorption in CQD particles, Frenkel pairs ([I_Pb−V_Pb]⁰) are created in CQDs, and some of these Frenkel pairs can dissociate into mobile charged defects, including the vacancy ([V_Pb]⁺¹) or interstitial ([I_Pb]⁻¹). When a strong electric field was applied across the CQD particles, the charged defects gained an electrokinetic energy by traversing the CQD particles. The mobile [I_Pb]⁻¹ can collide with [I_Pb−V_Pb]⁰, representing the first stage (Stage I). If these charged defects gain energy above E_th and collide with the neighboring CQD particles, intermediate states ([I_Pb − V_Pb − I_Pb]⁻¹) are formed (Stage II). These intermediated states then dissociate into one [V_Pb]⁺¹ and two [I_Pb] ⁻¹ (Stage III) in pursuit of lower energy states, which are energetically more stable. The multiplication of these [I_Pb]⁻¹ defects resulted in the creation of additional electron–hole pairs through integration with [I_Pb−V_Pb]⁰ (Stage I). A reduction in E_th was also observed with increasing d_DtoDs (Fig. 3c). This suggests that lower work functions, indicative of strong dipole moments in CQD solids with greater d_DtoDs, facilitate electron migration at the CQD–ligand interfaces (Figure S11 and Fig. 2e). This led to elevated solid-state reactions in regions adjacent to the charged defects, thereby triggering impact ionization at lower energy thresholds. Specifically, the E_th values were found to be 437, 224, and 189 meV for EthaneDT (3.86 nm), PropaneDT (4.06 nm), and PentaneDT (4.30 nm), respectively.

A charge density analysis revealed a significant accumulation of charge carriers at the CQD–ligand interfaces in all the thiol-treated CQDs (Figure S12). It was also confirmed that a shorter d_DtoD tended to increase the charge densities within the CQDs owing to the reduced electrokinetic energy required for electrons to migrate toward adjacent CQDs (Figure S13)⁴⁹. This suggests that regions with an elevated carrier concentration are spatially expanded on the CQD surfaces with shorter d_DtoDs values. These regions could serve as Auger recombination spots, in which the accelerated electrons consume energy and create phonons, thereby increasing E_th and reducing the charge multiplication factor in CQDs with shorter d_DtoDs values (Fig. 3d and Supplementary Note 5)⁵⁰.

However, an increase in d_DtoDs presents a challenge for the efficient charge transfer between neighboring CQDs. Specifically, an increased d_DtoD can facilitate radiative recombination before the accelerated electrons collide with the neighboring CQD particles (Fig. 2f). We calculated the hopping probabilities of CQDs with varied d_DtoDs values using a 1D multi-potential well model in COMSOL Multiphysics (Figure S13). We adjusted the width of the well to identify the impact of d_DtoDs on the CQDs, while the other parameters were obtained from DFT calculations (Supplementary Note 5 and Table S3). With increasing d_DtoDs, the hopping probability of the CQD particles decreased. This suggests that extended distances between neighboring CQDs can impede charge transfer, subsequently triggering radiative recombination.

We evaluated the current densities of the CQD-based IRPDs to analyze the effects of electron hopping and charge multiplication on infrared detection, depending on d_DtoDs within CQDs (Fig. 3e and Supplementary Note 5). EthaneDT-treated devices, with the shortest d_DtoDs of approximately 3.86 nm, exhibited a current density exceeding 100 mA/cm² under IR illumination (with a wavelength of 940 nm and a power of 24 µW). When d_DtoDs was increased to 4.06 nm (as in the PropaneDT-treated devices), the current density increased remarkably to 300 mA/cm². However, as the distance extended to 4.30 nm (as in the PentaneDT-treated devices), we noted a significant reduction in the current density (< 100 mA/cm²). The experimental results further validated the observed correlation between d_DtoDs and current density (Figure S14). The data suggest an optimal range of d_DtoDs that maximizes the current density while balancing the competing effects of charge multiplication, electron hopping, and recombination mechanisms.

Figure 3f schematically illustrates the effect of d_DtoDs resulting from the use of various ligands on the dynamics of charge transport and multiplication among the CQD particles. A shorter d_DtoD facilitated the rapid migration of the photogenerated electrons toward the neighboring CQDs (Fig. 3f, left). This also implies that the high rates of electron hopping increased the carrier densities and expanded the accumulated regions at the CQD–ligand interfaces (Figure S12). This led to an increase in phonon generation (i.e., Auger recombination) when the electrons passed through the accumulated regions in the thiol-treated CQD films. Because accelerated electrons dissipated their energy via phonon generation during the migration between adjacent CQDs, further acceleration was required to gain sufficient electrokinetic energy, thus requiring a higher E_th to trigger charge multiplication. On the other hand, as the neighboring CQDs moved further apart, the reduced rates of electron hopping suppressed the formation of Auger recombination spots, leading to a reduction in the thermalization loss in the CQDs (Fig. 3f, right). However, an increase in d_DtoD was accompanied by a reduction in the electron-hopping probability within the CQD-based IRPDs, increasing the radiative recombination (Fig. 2f).

Representative performance characterization of CQD-based IRPDs

Transient photocurrent (TPC) measurements revealed the beneficial impact of shorter distances between neighboring CQDs on charge transport within the CQD-based IRPD (Fig. 4a). The EthaneDT-treated device achieved a rising time (τ_r) of 0.85 µs, highlighting their enhanced charge extraction kinetics. Conversely, an increase in d_DtoDs between adjacent CQDs was observed to adversely increase τ_r, thereby undermining the charge transport capabilities of CQD-based IRPDs. Devices treated with PropaneDT, characterized by a mere 0.2 nm increase in d_DtoD relative to the EthaneDT-treated devices, exhibited a marginal increase in τ_r to 1 µs, indicating an inferior charge extraction efficiency. More interestingly, OctaneDT-treated devices with extended d_DtoDs of 5.48 nm exhibited a significant increase in τ_r exceeding 70-fold compared with the EthaneDT-treated devices. This suggests that the elongated d_DtoDs impede electron hopping between neighboring CQDs, resulting in decreased carrier mobility along with intensified PL emissions (Fig. 2f, Figure S4b, and Fig. 4b). The highest carrier mobility of 3.33×10^− 2 cm²/V⋅s was achieved when d_DtoD was minimized (i.e., EthaneDT) in CQD-based IRPDs among thiol-treated CQD devices. These findings underscore the intricate interdependency between d_DtoDs transitions and the diverse parameters that influence charge transport and recombination processes in these optoelectronic devices.

Under low-reverse-bias conditions (2 V), the optimization of the charge transport property in CQD-based IRPDs can help significantly reduce the noise current and improve the detectivity of CQD-based IRPDs without induced charge multiplication. As a result, EthaneDT-treated devices achieved a low noise current of 3 nA and an outstanding detectivity of 1.1×10¹³ Jones at 2 V. Conversely, longer d_DtoD caused a degradation in these key performance metrics. Specifically, OctaneDT-treated devices with extended d_DtoDs of 5.48 nm demonstrated a drastic reduction in the carrier mobility (1.01×10^− 5 cm²/V⋅s), leading to a lower operation speed and a reduced bandwidth of 20 kHz. Such changes in the device characteristics can increase the noise current to 2.06 µA⁵¹. This interplay resulted in a substantially compromised detectivity of 7.51×10⁶ Jones, which represents a six-order-of-magnitude decline relative to the EthaneDT-treated counterparts. These findings suggest that electron-hopping characteristics predominate as the main factor influencing the detectivity. Therefore, it is deduced that shorter d_DtoDs values between adjacent CQDs are advantageous for realizing superior detectivity in the absence of charge multiplication in CQD-based IRPDs.

Next, upon applying a reverse bias exceeding the breakdown voltage, we found that charge multiplication emerged as another prominent determinant of the detectivity across various CQD-based devices (Fig. 4c). The breakdown voltages in the current density–voltage curves and the critical thresholds in the operando PL spectra coincided with those of the CQD-based IRPDs (Table 1). Although short d_DtoDs between adjacent CQDs facilitate charge transport and detectivity, overreduction in these short d_DtoDs can bring additional challenges. Specifically, exceedingly short d_DtoDs values can facilitate thermal relaxation, thereby amplifying the Auger recombination under a high reverse bias. This increased Auger recombination causes a reduction in the current density in the high-reverse-bias region (> 15 V), consequently leading to a low multiplication gain (gain ≈ 14) and detectivity (5.6×10¹³ Jones) in the EthaneDT-treated devices (Figure S14).

Table 1

Comparative analysis of the threshold voltages derived from operando PL spectra (Fig. 2g) and breakdown voltages derived from the current density–voltage curves (Figure S13) in CQD-based IRPDs treated with EthaneDT, PropaneDT, PentaneDT, and OctaneDT.
	EthaneDT	PropaneDT	PentaneDT	OctaneDT
Threshold voltage [V]	13	21	24.5	38.5
Breakdown voltage [V]	12.7	21.3	25.2	38

With the increase in d_DtoDs, the observed reduction in the photocurrent is mitigated, thereby allowing for an enhanced multiplication gain in CQD-based devices under the application of a reverse bias exceeding 24 V. This facilitated the acquisition of greater kinetic energy by the photoelectrons, thereby achieving a high multiplication gain of 85 in the PropaneDT-treated devices. However, when d_DtoDs exceeded 4.06 nm, the inferior electron hopping characteristics led to the quenching of accelerated electrons via radiative recombination before electron collision, as illustrated in Figs. 2f and S9. This limitation implies that only a limited number of accelerated electrons contribute to the charge multiplication process, leading to a lower multiplication gain and detectivity in the PentaneDT- and OctaneDT-treated devices than in the PropaneDT-treated devices. Consequently, a representative detectivity of 1.4×10¹⁴ Jones could be achieved in the PropaneDT-treated devices (Fig. 4d). Furthermore, our CQD-based devices exhibited a substantial linear dynamic range of 124 dB, indicating efficient operation at light intensities ranging from 10 µW to 300 mW (Fig. 4e). The CQD-based devices also demonstrated a high-speed performance across a broad frequency spectrum ranging from 10 to 200 kHz (Fig. 4f).

In this study, we devised an innovative architecture that promotes electrokinetically pumped charge multiplication within the CQD layer, thereby enabling exceptional levels of detectivity in CQD-based IRPDs with thiol-based ligands. Our findings revealed that maintaining a CQD layer thickness in excess of 540 nm is crucial for maximizing charge multiplication while suppressing noise currents due to electron tunneling. Although the elongation of the distances between adjacent CQDs enhances charge multiplication, it adversely affects the probability of electron hopping in thiol-treated CQD devices. Consequently, we achieved a notable detectivity of 1.4×10¹⁴ Jones in CQD-based IRPDs incorporated with PropaneDT ligands, corresponding to a d_DtoD value of 4.06 nm. Furthermore, our CQD-based devices exhibited a wide linear dynamic range of 124 dB and a rapid response time up to an operating frequency of 200 kHz without the need for additional treatments. This economical and charge-multiplication-capable CQD-based IRPD architecture has significant potential for a range of applications, particularly in emerging sensors that require high-performance IRPDs.

Materials

All the devices were deposited on pre-patterned indium tin oxide (ITO) substrates with a sheet resistance of 22 Ω/sq, purchased from JM International Co., Ltd. Organic solvents, such as octane, acetonitrile, and methanol, were purchased from Sigma-Aldrich. All the thiol-based ligands, including 1,2-ethanedithiol (EthaneDT), 1,3-propanedithiol (PropaneDT), 1,5-pentanedithiol (PentaneDT), and 1,8-octanedithiol (OctaneDT), were purchased from Sigma-Aldrich. PbS CQD capped with oleic acid and sol-gel ZnO solutions were synthesized as previously described⁵².

Device fabrication

Prepatterned ITO-coated glass substrates were cleaned with detergents, acetone, and 2-propanol (IPA), followed by a 10 min O₂ plasma treatment before film deposition. The sol-gel ZnO layer was deposited by spin coating at 2000 rpm for 30 s. A 30 nm-thick ZnO layer was formed after thermal annealing at 200°C for 30 min. Utilizing a layer-by-layer deposition technique, 13 layers of thiol-passivated PbS CQD layers were sequentially coated onto the ZnO layer via exchange processes: For each layer, approximately 25 µl of a PbS CQD solution (50 mg∙ml^− 1 in octane) was spin-coated on the ZnO layer at 2500 rpm for 15 s. Ligand exchange was executed using thiol-based ligands, including EthaneDT, PropaneDT, PentaneDT, and OctaneDT solution (0.02% w/v in acetonitrile). This ligand solution was applied to the CQDs film for 30 s, followed by rinsing thrice with acetonitrile. The coated CQD films were air-dried overnight. Finally, Au (150 nm) was deposited onto the active film via thermal evaporation using a shadow mask.

Characterization

The absorption spectra of the CQD films subjected to various thiol treatments were measured using a UV/Vis/NIR spectrophotometer (Shimadzu, UV3600Plus). The current density–voltage (J–V) characteristics were measured from 1 to − 50 V using a photovoltaic simulator (K201 LAB55, McScience) at an irradiance of 24 µW∙cm^− 2 from a 940 nm laser diode. The light intensity was calibrated using an optical power meter (SANWA, Laser power meter LP10). The illumination area of each device was 0.15 cm². High-resolution transmission electron microscopy (HR-TEM) and scanning TEM-energy dispersive spectroscopy (STEM-EDS) were used for the morphological analysis of the samples. Cross-sectional images of the overall devices were captured using a focused ion beam (FIB, Helios 450 F1) and field-emission TEM operating at 300 keV. Transient photocurrent (TPC) and transient photovoltage (TPV) spectra of the devices were measured using a parameter analyzer (McScience, T4000), under an irradiance of 24 µWžcm^− 2. During the TPC measurements, a light pulse of 15 µs duration was applied under ambient conditions (298 K, 30% humidity). The data presented are the averages of 16 independent measurements. Reproducible on/off switching current measurements were performed using a probe station (M5VC, MSTECH) under a reverse bias of 25 V and a light source of a 940 nm laser diode (24 µWžcm^− 2). The chemical binding states and valence band-edge states of the CQDs were analyzed using X-ray photoelectron spectroscopy (XPS; Axis-Supra, Kratos Analytical, United Kingdom) with Al Kα source (1486.6 eV). Stacking structure analyses were conducted using grazing incidence small angle X-ray scattering (GISAXS; Nanopics, Rigaku, Japan).

Data availability

The authors declare that the primary data underpinning the findings of this study are available within the article and its Supplementary Information. Additional data are available from the authors upon request.

Acknowledgments

This study was supported by the National Research Foundation of Korea (NRF) grant funded by the Korean Government (MSIT) (NRF-2022M3H4A1A03076626).

Author contributions

B. K., S. Y. L., and H. K. contributed equally to this study. B.K., S. Y. L., H.K., and J.-Y. L. developed the concept and prepared the manuscript. B.K. and J.L. characterized the electrical properties of each film. J.L. and S.J.C optimized the structure of each device. S. Y. L. characterized the chemical properties in each film. H.K. and H.S. developed the density functional theory calculation. M.L. constructed the 2D device models. J.-Y.L. supervised the study. All the authors discussed the results and commented on the manuscript.

Additional information

Competing financial interests

The authors declare no competing financial interests.

Chen, D. et al. Photon-trapping-enhanced avalanche photodiodes for mid-infrared applications. Nature Photonics 17, 594-600 (2023).
Kim, J. et al. A skin-like two-dimensionally pixelized full-color quantum dot photodetector. Science Advances 5, 1-12 (2019).
Gong, X. et al. High-Detectivity Polymer Photodetectors with Spectral Response from 300 nm to 1450 nm. Science 325, 1665-1667 (2009).
Zhang, X., Kwon, K., Henriksson, J., Luo, J. & Wu, M. C. A large-scale microelectromechanical-systems-based silicon photonics LiDAR. Nature 603, 253-258 (2022).
Horiuchi, N. Infrared applications. Nature Photonics 13, 376-377 (2019).
Zhang, T. et al. Six-arm Stellat Dendritic-PbS Flexible Infrared Photodetector for Intelligent Healthcare Monitoring. Advanced Materials Technologies 7, 2200250 (2022).
Wang, M. et al. Silicon-Based Intermediate-Band Infrared Photodetector Realized by Te Hyperdoping. Adv. Opt. Mater. 9, 2001546 (2021).
Lischke, S. et al. Ultra-fast germanium photodiode with 3-dB bandwidth of 265 GHz. Nature Photonics 15, 925-931 (2021).
Jones, A. H., March, S. D., Bank, S. R. & Campbell, J. C. Low-noise high-temperature AlInAsSb/GaSb avalanche photodiodes for 2-μm applications. Nature Photonics 14, 559-563 (2020).
Li, N. et al. Solution-processable infrared photodetectors: Materials, device physics, and applications. Materials Science and Engineering: R: Reports 146, 100643 (2021).
Shen, L., Pun, E. Y. B. & Ho, J. C. Recent developments in III–V semiconducting nanowires for high-performance photodetectors. Materials Chemistry Frontiers 1, 630-645 (2017).
García de Arquer, F. P. et al. Semiconductor quantum dots: Technological progress and future challenges. Science 373, 1-14 (2021).
Norris, D. J. Multispectral quantum-dot photodetectors. Nature Photonics 13, 230-232 (2019).
Ackerman, M. M. Bringing Colloidal Quantum Dots to Detector Technologies. Information Display 36, 19-23 (2020).
Septianto, R. D. et al. On-demand tuning of charge accumulation and carrier mobility in quantum dot solids for electron transport and energy storage devices. NPG Asia Mater. 12, 33 (2020).
Hartley, C. L. & Dempsey, J. L. Revealing the Molecular Identity of Defect Sites on PbS Quantum Dot Surfaces with Redox-Active Chemical Probes. Chem. Mat. 33, 2655-2665 (2021).
Beygi, H., Sajjadi, S. A., Babakhani, A., Young, J. F. & van Veggel, F. C. J. M. Surface chemistry of as-synthesized and air-oxidized PbS quantum dots. Applied Surface Science 457, 1-10 (2018).
Vafaie, M. et al. Colloidal quantum dot photodetectors with 10-ns response time and 80% quantum efficiency at 1,550 nm. Matter 4, 1042-1053 (2021).
Seo, H. et al. Cationic Molecular Metal Chalcogenide Ligand-Passivated Colloidal Quantum Dots and Their Application to Suppressed Dark-Current Near-Infrared Photodetectors. Advanced Materials Technologies 8, 2201864 (2023).
Bederak, D. et al. Comparing Halide Ligands in PbS Colloidal Quantum Dots for Field-Effect Transistors and Solar Cells. ACS Applied Nano Materials 1, 6882-6889 (2018).
Nakazawa, N. et al. The interparticle distance limit for multiple exciton dissociation in PbS quantum dot solid films. Nanoscale Horizons 4, 445-451 (2019).
Sukhovatkin, V., Hinds, S., Brzozowski, L. & Sargent, E. H. Colloidal Quantum-Dot Photodetectors Exploiting Multiexciton Generation. Science 324, 1542-1544 (2009).
Kim, S. J., Kim, W. J., Sahoo, Y., Cartwright, A. N. & Prasad, P. N. Multiple exciton generation and electrical extraction from a PbSe quantum dot photoconductor. Applied Physics Letters 92, 031107 (2008).
Wang, S. et al. Analysis of breakdown probabilities in avalanche photodiodes using a history-dependent analytical model. Applied Physics Letters 82, 1971-1973 (2003).
Lee, S. et al. Engineering of impact ionization characteristics in In0.53Ga0.47As/Al0.48In0.52As superlattice avalanche photodiodes on InP substrate. Scientific Reports 10, 16735 (2020).
Yi, X. et al. Extremely low excess noise and high sensitivity AlAs0.56Sb0.44 avalanche photodiodes. Nature Photonics 13, 683-686 (2019).
Chen, J. et al. High-performance HgCdTe avalanche photodetector enabled with suppression of band-to-band tunneling effect in mid-wavelength infrared. npj Quantum Materials 6, 103 (2021).
Bai, P. et al. Broadband and photovoltaic THz/IR response in the GaAs-based ratchet photodetector. Science Advances 8, 1-9 (2022).
Lu, Z. et al. Geiger-Mode Operation of Ge-on-Si Avalanche Photodiodes. IEEE Journal of Quantum Electronics 47, 731-735 (2011).
Kim, H., Choi, S. & Choi, B. J. Influence of AlN and GaN Pulse Ratios in Thermal Atomic Layer Deposited AlGaN on the Electrical Properties of AlGaN/GaN Schottky Diodes. Coatings 10, 489 (2020).
Chen, B. et al. Low Dark Current High Gain InAs Quantum Dot Avalanche Photodiodes Monolithically Grown on Si. ACS Photonics 7, 528-533 (2020).
Zhang, B. Y., Yang, C., Liu, W. F. & Liu, A. M. Origin of breakdown mechanism in multicrystalline silicon solar cells. Applied Physics Letters 101, 093903 (2012).
Aftab, S. et al. Switching photodiodes based on (2D/3D) PdSe2/Si heterojunctions with a broadband spectral response. Journal of Materials Chemistry C 9, 3998-4007 (2021).
Guo, B. et al. Temperature Dependence of Avalanche Breakdown of AlGaAsSb and AlInAsSb Avalanche Photodiodes. Journal of Lightwave Technology 40, 5934-5942 (2022).
Mukherjee, A. et al. Vertical Architecture Solution-Processed Quantum Dot Photodetectors with Amorphous Selenium Hole Transport Layer. ACS Photonics 10, 134-146 (2023).
Reddy, P. et al. Study on avalanche breakdown and Poole–Frenkel emission in Al-rich AlGaN grown on single crystal AlN. Applied Physics Letters 119, 182104 (2021).
Liu, M. et al. Lattice anchoring stabilizes solution-processed semiconductors. Nature 570, 96-101 (2019).
Weidman, M. C., Smilgies, D.-M. & Tisdale, W. A. Kinetics of the self-assembly of nanocrystal superlattices measured by real-time in situ X-ray scattering. Nature Materials 15, 775-781 (2016).
Pinna, J. et al. Approaching Bulk Mobility in PbSe Colloidal Quantum Dots 3D Superlattices. Advanced Materials 35, 2207364 (2023).
Ip, A. H. et al. Hybrid passivated colloidal quantum dot solids. Nature Nanotechnology 7, 577-582 (2012).
Hwang, G. W. et al. Identifying and Eliminating Emissive Sub-bandgap States in Thin Films of PbS Nanocrystals. Advanced Materials 27, 4481-4486 (2015).
Liu, M. et al. Controlled Steric Hindrance Enables Efficient Ligand Exchange for Stable, Infrared-Bandgap Quantum Dot Inks. ACS Energy Letters 4, 1225-1230 (2019).
Morris, T. & Zubkov, T. Steric effects of carboxylic capping ligands on the growth of the CdSe quantum dots. Colloids and Surfaces A: Physicochemical and Engineering Aspects 443, 439-449 (2014).
Choi, M. et al. Tailored Band Edge Positions by Fractional Ligand Replacement of Nonconductive Colloidal Quantum Dot Films. The Journal of Physical Chemistry C 127, 4825-4832 (2023).
Bultinck, P., Goeminne, A. & Van de Vondel, D. Ab initio conformational analysis of 1,3-propanedithiol. Journal of Molecular Structure: THEOCHEM 360, 119-125 (1996).
Dibenedetto, C. N. et al. Coupling in quantum dot molecular hetero-assemblies. Materials Research Bulletin 146, 111578 (2022).
Yang, S., Prendergast, D. & Neaton, J. B. Tuning Semiconductor Band Edge Energies for Solar Photocatalysis via Surface Ligand Passivation. Nano Letters 12, 383-388 (2012).
Xu, Z. et al. Discovery of ionic impact ionization (I3) in perovskites triggered by a single photon. arXiv preprint arXiv:1906.02475 (2019).
LaRue, J. et al. Vibrationally promoted electron emission at a metal surface: electron kinetic energy distributions. Physical Chemistry Chemical Physics 13, 97-99 (2011).
Gao, Y., Sandeep, C. S. S., Schins, J. M., Houtepen, A. J. & Siebbeles, L. D. A. Disorder strongly enhances Auger recombination in conductive quantum-dot solids. Nature Communications 4, 2329 (2013).
Chow, P. C. Y. & Someya, T. Organic Photodetectors for Next-Generation Wearable Electronics. Advanced Materials 32, 1902045 (2020).
Kim, B., Baek, S.-W., Kim, C., Kim, J. & Lee, J.-Y. Mediating Colloidal Quantum Dot/Organic Semiconductor Interfaces for Efficient Hybrid Solar Cells. Advanced Energy Materials 12, 2102689 (2022).

Supplementary Notes 1 to 5, Supplementary Figures, and Supplementary Tables are not available with this version.

The authors declare no competing interests.

Tableofcontentsentry.docx

Download PDF

Version 1

posted

You are reading this latest preprint version

Ultrahigh-gain colloidal quantum dot infrared photodetectors: Unraveling the potential of electro-kinetically pumped charge multiplication

Status:

Version 1

Abstract

Figures

Introduction

Results and discussion

CQD layer thickness effect on charge multiplication

Characterization of the distances between CQD nanoparticles

Theoretical calculations for CQD nanoparticles depending on dot-to-dot distances

Representative performance characterization of CQD-based IRPDs

Conclusions

Methods [or Experimental section]

Declarations

References

Supplementary Information

Additional Declarations

Supplementary Files

Status:

Version 1