Color-selective photodetection from intermediate colloidal quantum dots buried in amorphous-oxide semiconductors

We report color-selective photodetection from intermediate, monolayered, quantum dots buried in between amorphous-oxide semiconductors. The proposed active channel in phototransistors is a hybrid configuration of oxide-quantum dot-oxide layers, where the gate-tunable electrical property of silicon-doped, indium-zinc-oxide layers is incorporated with the color-selective properties of quantum dots. A remarkably high detectivity (8.1 × 1013 Jones) is obtained, along with three major findings: fast charge separation in monolayered quantum dots; efficient charge transport through high-mobility oxide layers (20 cm2 V−1 s−1); and gate-tunable drain-current modulation. Particularly, the fast charge separation rate of 3.3 ns−1 measured with time-resolved photoluminescence is attributed to the intermediate quantum dots buried in oxide layers. These results facilitate the realization of efficient color-selective detection exhibiting a photoconductive gain of 107, obtained using a room-temperature deposition of oxide layers and a solution process of quantum dots. This work offers promising opportunities in emerging applications for color detection with sensitivity, transparency, and flexibility.


Supplementary Method 2. Calculation of charge separation rate in OQO devices
The kinetics of the excited QD population ( ( )) can be simply expressed by where is a coefficient and is a decay rate. This can be resolved into the dynamics of a photocarrier decay rate ( ) in QDs and a charge separation rate ( ) from QDs to SIZO layers in OQO configuration as follows: Here, we can easily obtain the charge separation time ( = 1/ ) expressed by = +

(Supplementary Equation 3)
where is 0.29 ns and is 7.7 ns from the fitting curves as indicated by green and black solid lines in Fig. 3a, respectively. Therefore, we can extract the charge separation time

Supplementary Note 1. Comparison of interface area between QD and SIZO
As shown in Supplementary Fig. 5, the contact interfacial area between QDs and oxides in OQO structure (Supplementary Fig. 5a) should be much larger (more than 2 times) than QO configuration ( Supplementary Fig. 5b where is the radius of QDs and is the mean free path in bulk 30 . Considering CdSe, becomes only 2.2 nm, because is about 20 nm in bulk and is around 2.5 nm. This is comparable to the size of QDs so that photoexcited charges in the middle of QD layers are going to experience serious collisions even in the 3Q-OQO structure. Thirdly, the effective field strength given by the gate voltage becomes weak due to the increase of gap distance between source and gate electrodes. Thus, light power we used (5 ) is still far from the absorption saturation regime.

Supplementary Note 3. Calculation of saturable light power for mono-layer
In the QO phototransistor in [Ref 7], the absorption saturation might be similar to the above calculated values. But in the QO phototransistor, the saturation of photocurrent occurred at the light power intensity of 300 W/cm 2 , which mean the charge separation ( QO ) is very slow, compared with our OQO configuration. We investigated this value as 18.1 ns (1/ QO = 1/ QD + 1/ S, QO ), where QO = . ns and QD = . ns.
In contrast, the charge saturation in our phototransistors was not observed even at high intensity of light power (255-mW/cm 2 ) and this is one of the advantages of fast charge separation in OQO configuration.

Supplementary Note 4. Inherent photoconductive gain
We employ the new concept of G inher , including net absorption in the OQO channels, by eliminating the device R light and T light . The red curve shown in Supplementary Fig. 10a is the net absorptions of the 1RQ-OQO channel. The absorbance at 487 nm is 0.0008, which indicates 0.02% light absorption. Thus, G inher is 500 times larger than G meas , as shown by the blue and red curves in Fig. 3c, respectively.

Supplementary Note 5. Photoconductive gain using photocurrent-response decay times
Supplementary Fig. 11 shows the temporal response of the monolayered-red QD OQO and G max ) are marked in Fig. 3c.

Supplementary Note 6. Transparencies of SIZO and QO films
The transparencies of the SIZO-only and the QD-on-SIZO films are shown in Supplementary Fig. 12. The transparency depends on the QD layer thickness. The 6-layeredred-QD (6RQ)-on-SIZO film exhibits the lowest transparency of 90% at 400-nm wavelength.
In contrast, the films with monolayered QDs on SIZO for the different RGB colors exhibit higher transparencies of more than 95% at the same wavelength.

Supplementary Note 7. Johnson noise calculation
The Johnson noise is 4kTf /R, where k is the Boltzmann constant, T is the temperature, f is the noise bandwidth, and R is the detector resistance under dark conditions. In the subthreshold regime, A (V G = -15V), R is ca. 10 12 ; therefore, the Johnson noise is 1.6 ⅹ 10 -32 A 2 Hz -1 . In the ohmic (B) and near-saturation regimes (C), the Johnson noise becomes 1.6 ⅹ 10 -29 and 1.6 ⅹ 10 -26 A 2 Hz -1 , respectively.

Supplementary Note 8. Effect of field screening by QDs
In our OQO phototransistors, the thickness of QD layers is less than a few ten of nanometers, e.g., 30 nm for 6-layered QD films. Thus, the field screening effect of QDs is considered small that top SIZO layers play a role as a charge transport layer. For example, we prepared OQ phototransistors with a 40-nm oxide layer on top of 6-layered QDs and investigated whether the top oxide layer can support charge flows, as depicted in Supplementary Fig. 14. It is evidently shown that even the oxide channel layer on top of 30nm thick, 6-layered QDs can be successfully gated by V G .