Spatially Resolved Photo-Response of a Carbon Nanotube/Si Photodetector

Abstract

Photodetectors based on vertical multi-walled carbon nanotube (MWCNT) film-Si heterojunctions are realized by growing MWCNTs on n-type Si substrates with a top surface covered by Si₃N₄ layers. Spatially resolved photocurrent measurements reveal that higher photo detection is achieved in regions with thinner MWCNT film, where nearly 100% external quantum efficiency is achieved. Hence, we propose a simple method based on the use of scotch tape with which to tune the thickness and density of as-grown MWCNT film and enhance device photo-response.

Keywords: carbon nanotubes; heterostructure; photoconductivity; photodetector; photodiode; quantum efficiency; silicon heterostructure.

Figure 1

MWCNT film-thinning process: (a) device with pristine MWCNT film, (b) scotch tape adhered to half area of the device, and (c) tape peeled off from the device. (d) The thickness of the MWCNT film along the lines in (c). (e) The layout of the device and the setup used for the photo-response characterization as a function of the light spot position.

Figure 2

(a) I–V characteristic of the device in the dark and under illumination by a 380 nm LED. V_ph is the voltage used to estimate the photocurrent. (b) Fowler–Nordheim plot of the blue curve in (a). (c) Band structure of the MIS junction formed between the nanotubes, the silicon nitride, and the n-doped silicon at equilibrium, forward, and reverse bias. In the reverse bias condition, the photogenerated holes inside the silicon can tunnel through the triangular-shaped barrier of the silicon nitride, giving rise to a photocurrent.

Figure 3

(a) Quantum efficiency map of a device (with pristine MWCNT film) that shows efficiency inhomogeneities across its surfaces. The device presents an average QE of 30%, but near the top and bottom edges, the QE increases to nearly 100%. (b) SEM images of the top-right corner of the MWCNT film. The film presents two types of edges: the top edge, which decreases gradually, and the right edge, which is sharper and akin to a step.

Figure 4

(a) SEM image of the device tilted by a 15-degree angle from the surface plane after the thinning process. (b) Optical image of the device after the thinning process. Quantum efficiency map of the device before (c) and after (d) MWCNT removal. The black lines mark the real dimensions of the substrate and metallic pads. Both maps were acquired using the right pad as an electrical contact.

Figure 5

(a) Average QE of the device at the plateau and associated standard deviation (error bars) as a function of the wavelength. Red dots represent the values of the QE obtained on the pristine portion of the film, while the blue dots represent the values from the thinned portion. (b) Comparison of the I-V curves acquired using the left or right pads as electric contacts (indicated by the letters L and R), in the dark and under the light of a 650 nm laser at 250 µW (with light spot focused on the thinned or pristine film).