The Comparison of Colloidal PbS QD Photoconductors and Hybrid Phototransistors

Abstract

The simplicity in the fabrication of photoconductors makes them a valuable choice to investigate optoelectronic properties of colloidal quantum dot (cQD) films. Lateral photoconductors generally require a large size, in the mm2, and are limited in operation speed due to the presence of trapping sites. In contrast, hybrid phototransistors are fabricated in the um2 scale and benefit from such trapping sites, allowing the measurement of low light levels in the nW/cm2. The question, however, arises whether high responsivity values are required for the detection of low light levels or the compatible detectivity of photoconductors is sufficient. Here, we directly compare photoconductors and hybrid phototransistors with an identical EDT-treated PbS cQD film. We highlight that a comparable D* is not enough for the purpose of measuring low light levels, as the resulting photocurrents need to be readily accessible. Furthermore, we also showcase temperature-activated photocurrent dynamics resulting in a negative photocurrent (NPC) effect. This NPC simultaneously improves the frequency bandwidth and photocurrent, enabling operation speeds up to 100 kHz.

Figures (3)

• Figure 1: IFP. (a) IFP consisting of 30 gaps with a L of 10 and a W of 500, fabricated on a p-Si/SiO2 substrate. A $\sim$170nm PbS cQD film was spin-coated subsequently. A source-drain voltage $V_{DS}$ and a back-gate voltage $V_G$ were applied, while the source-drain current $I_{DS}$ was measured. A chopper modulated the incoming light at a frequency $f_{chop}$, causing a photocurrent $I_{ph, AC}$ on top of $I_{DS}$. $I_{ph, AC}$ was extracted by a lock-in technique over the voltage drop across a shunt resistor $R_{shunt}$. The inset shows the structure before cQD deposition. (b) EDT treated PbS cQD film energy diagram. A p-type charge transport behavior follows from a higher hole mobility $\mu_h$ than an electron mobility $\mu_e$. The photoconduction of the IFP is expressed as a vertical shift of the source-drain current upon illumination. (c) $V_G$-dependent $I_{DS}$ measured in the dark for three different temperatures. The curves show two loops (forward and backward sweeps). (d) Comparison of $V_G$-dependent photocurrents measured at constant illumination, $I_{ph,const} = I_{DS,light} - I_{DS,dark}$ (solid lines), or AC modulated, $I_{ph,AC}$ (dotted lines), at 6Hz. Curves are shown for the same temperatures as in (c). For simplicity, the constant illuminated curves are shown in the forward sweep direction only (e) AC modulated responsivity $R_{AC}=I_{ph,AC}/P_{in}$, mapped over temperature $T$ (80 to 300K) and $V_G$. All the measurements upon light exposure were performed at the first excitonic peak (1550nm wavelength) of the $\sim$6nm cQD, and an irradiance ($\mathds{I}_{in}=P_{in}/A$) of 120μ W cm^2. $V_{DS}$ was kept at 1V for all the measurements.
• Figure 2: HP. (a) CVD graphene with a channel of $L\times W= 20\times1$ ^2 was patterned on a p-Si/SiO2 substrate. The measurements were performed as described above. Inset shows the graphene channel before the layer-by-layer spin-coating of PbS cQD. (b) Energy diagram for the graphene-cQD interface. (i) The Schottky-like junction supports a higher hole than the electron-transfer rate across the graphene-cQD film interface. This results in a negatively charged cQD film, top-gating the graphene channel (ii). The photogating effect is expressed as a horizontal shift of the $I_{DS}-V_G$ curves upon light illumination. (c) $V_G$-dependent channel current $I_{DS}$, measured in dark and for three different temperatures. After cooling, the charge carrier mobility of the channel increased from 2000 to about 2500cm^2 Vs. On the left and right of the charge neutrality point (CNP), the charge carriers in graphene are holes (p-doped) or electrons (n-doped), respectively. (d) Photocurrent comparison at constant illumination, $I_{ph,const}$ (solid lines), and 6Hz AC-light-modulated, $I_{ph,AC}$ (dotted lines), condition. The currents are shown for the same temperatures as in (c). (e) AC-light-modulated responsivity $R_{AC}$, mapped for $V_G$ and a temperature range between 80 to 300K. A dashed line highlights the CNP. All the light measurements were performed at a wavelength $\lambda$ of 1550nm and irradiance ($\mathds{I}_{in}=P_{in}/A$) of 120μ W cm^2. $V_{DS}$ was kept at 1V.
• Figure 3: Photocurrent dynamics of IFP and HP. (a) Transient photocurrent measurement setup. A trans-impedance amplifier was used to apply a source-drain bias $V_{DS}$ and measure the current through the photodetectors. Before subsequent sampling, the signal was low pass filtered at 1kHz. The illumination was controlled by a shutter (10ms rise/fall time). (b) transient photocurrents for light on (orange shading) and off (white) of the IFP, measured at a positive gate voltage and a temperature of 300 (red) or 80K (blue). Two time constants $\tau$ were extracted from the photoresponse upon illumination, where the fast contribution $\tau_{1, on}$ reduces by at least one order of magnitude upon cooling. A NPC spike occurs upon illumination of the device at 80K. The enlargement of the region of interest (gray shading) illustrates a formation of a Schottky-like barrier between the cQD film and gold electrodes. Illumination at the interface in (i) leads to a fast escape of photoexcited electrons opposing the main current (NPC). The illumination of the rest of the detector (ii) leads to a photocurrent contribution in bias direction and re-establishes the PPC eventually. The oscillations in the signal are contributions from the line frequency. (c) transient photocurrent of the HP showing NPC upon illumination and turning off the light. The NPC is modeled by a barrier formation at the interface altering the initial band alignment (i). While the photoexcited holes are trapped, a fast electron transfer to graphene renders the cQD-photogate positive and causes an NPC spike (ii). The increased hole concentration readjusts the barrier, allowing the cQD film to deplete the holes and reestablishes the PPC (iii). Upon turning off the light, electrons transfer back into the cQD film, causing another NPC spike (iv) before reaching the initial dark current value (v). (d) Photocurrent frequency response (AC-modulated light measured with lock-in technique). NPC causing a bandwidth improvement $\Delta f$. This leads to a sacrifice of the photocurrent by $\Delta I_{ph}$ for the IFP (bottom). The HP (top) shows a simultaneous $\Delta I_{ph}$ and $\Delta f$ improvement.