Ultrahigh Performance C60 Nanorod Large Area Flexible Photoconductor Devices via Ultralow Organic and Inorganic Photodoping

One dimensional single-crystal nanorods of C60 possess unique optoelectronic properties including high electron mobility, high photosensitivity and an excellent electron accepting nature. In addition, their rapid large scale synthesis at room temperature makes these organic semiconducting nanorods highly attractive for advanced optoelectronic device applications. Here, we report low-cost large-area flexible photoconductor devices fabricated using C60 nanorods. We demonstrate that the photosensitivity of the C60 nanorods can be enhanced ~400-fold via an ultralow photodoping mechanism. The photodoped devices offer broadband UV-vis-NIR spectral tuneability, exhibit a detectivitiy >109 Jones, an external quantum efficiency of ~100%, a linear dynamic range of 80 dB, a rise time 60 µs and the ability to measure ac signals up to ~250 kHz. These figures of merit combined are among the highest reported for one dimensional organic and inorganic large-area planar photoconductors and are competitive with commercially available inorganic photoconductors and photoconductive cells. With the additional processing benefits providing compatibility with large-area flexible platforms, these devices represent significant advances and make C60 nanorods a promising candidate for advanced photodetector technologies.

R6G is a well-known n-type organic dye and has been previously employed as a photosensitizer to enhance the photosensitivity of devices. [1][2][3] The absorption spectra of R6G in methanol is shown in figure S1a along with its relative energy level alignment with C60 in figure S1b. C60 nanorod devices sensitized with R6G (0.5mg/ml in methanol) were also found to increase the responsivity of the device. A comparison of spectral responsivity of the C60 and R6G photodoped C60 device is shown in figure S2b. The photocurrent is seen to be enhanced particularly in the spectral region above ~575 nm, where C60 only absorbs, demonstrating the benefits of the photodoping provided by R6G filling C60 traps ( Figure S2a). A contribution to the photocurrent from R6G can also be seen in the 500 nm to 550 nm with a shoulder in the responsivity spectrum at ~520 nm corresponding to the peak of the R6G absorption. Figure S1. Absorption and energy level alignment (a) Absorption spectrum of R6G in methanol and TCNQ in acetonitrile (b) Relative energy level alignment between C60 and variety of materials used in the study to photodope C60 nanorods. Unlike other dopant materials used, TCNQ doesn't form a type-II heterojunction with C60, thus photoinduced electron transfer from TCNQ to C60 is energetically forbidden. The energy level values for R6G and TCNQ are taken from reference [2] and [4] respectively.
The transient response of C60 rod only devices and R6G photodoped devices obtained using pulsed (~8 ns pulse width at 21 Hz repetition rate) 520 nm illumination are shown in S2d. It can be seen that the photocurrent obtained from the C60 rods only device decays faster as compared to R6G photodoped devices indicating a direct dependence of the photocurrent generation process on the photodopant. This, along with the enhanced spectral response, precludes the increase in photocurrent arising solely from the dopant reducing contact resistance between C60 rods for example. Figure S2. Photodoping C60 nanorods with rhodamine 6G (a) Normalised photocurrent (at 460 nm) of C60 nanorods only and R6G photodoped devices. (b) Spectral responsivity of C60 nanorod and R6G photodoped C60 nanorod devices measured at an applied electric field strength of 1V/µm. (c) J-V characteristic of the R6G photodoped C60 nanorod device. (d) Normalized transient photocurrent response of C60 nanorod device before and after photodoping with R6G. Photocurrent in the R6G photodoped decays slowly relative C60 rods indicating the dopant hole trapping process. Inset to (d) is shown the full transient response of the R6G photodoped device. Transient decays were obtained using a ~8 ns pulse at 21 Hz repetition rate exciting at ~520 nm (1.2 mW) with the device biased with a 1V/µm electric field.
Conversely when TCNQ was used as the 'photodopant' a reduction in device photocurrent was observed across the entire spectral region ( Figure S3a). Inspection of the relative energy level alignment between C60 and TCNQ shows that the TCNQ lowest unoccupied molecular orbital (LUMO) lies just above the mid-point of the C60 LUMO and highest occupied molecular orbital (HOMO). As such excitation of TCNQ will not then allow electron transfer into C60 trap states located below its LUMO and therefore we would not expect any increase in photocurrent as observed for the other photodopants utilised. Furthermore, it can be seen that in addition to TCNQ absorbing photons, which therefore will not contribute to the photocurrent, it is possible for excited electrons within C60 to relax into the unexcited TCNQ LUMO. These effects result in the reduction of the photocurrent as observed and further support the proposed mechanism behind the operation of the devices reported within. We note that devices formed from TCNQ alone show a weak photoresponse matching the TCNQ absorption spectrum ( Figure S3b).

Effect of oxygen on C60 device performance
C60 crystals intrinsically show good conductivity however this is significantly reduced upon exposure to oxygen. For example, C60 crystals prepared via solvent evaporation and C60 nanorods prepared via LLIP method have shown very high electron mobilities with reported values of 11 cm 2 V -1 S -1 and 1 cm 2 V -1 S -1 respectively, when measured in nitrogen environment. 5,6 To demonstrate this, we fabricated a C60 nanorod device in nitrogen and performed I-V measurements under nitrogen for ~ 30 minutes before exposing the same device to air ( Figure   S4). The device exhibits an initially high conductivity in nitrogen which over ~30 minutes does not degrade significantly. Upon exposure to air a significant decrease in conductivity is observed followed by a further decrease until after ~2.5 hours the current stabilized. We note that all of the devices reported in this study were fabricated in air under room light and therefore it is likely that this effect is the origin of the traps which photodoping process fills. Figure S4. Effect of air exposure on conductivity of C60 nanorod device. Dark I-V characteristics of a typical C60 nanorod device fabricated and operated in nitrogen prior to exposure to air.

HRTEM Studies of Photodoped C60 Rods
In Figure S5

Methods for enhancing photocurrent
Our results clearly show that the conductivity of C60 nanorods decreases drastically with exposure to oxygen, which is due to the creation of trap states that hamper the hopping transport mechanism within the C60 nanorods ( figure S4). We further demonstrate that the hopping process can be made efficient in C60 rods by filling electron traps states. As discussed in the manuscript this can be done by increasing the charge carrier concentration in C60 in three different ways: (i) by increasing the applied bias; (ii) by increasing the optical intensity; and (iii) by using photodopants (including PbS and CdSe nanocrystals, P3HT and rhodamine 6G).
In Figure S6 we demonstrate these methods using PbS NCs as the photodopant material. In Figure S6a increasing the applied electric field is seen to lead to a direct increase in photocurrent. Similarly, increasing the incident light intensity at a fixed bias also increases the photocurrent ( Figure S6b). Finally, figure S6c shows directly the effect of increasing the PbS NC concentration on increasing the photocurrent. The increase in the dopant concentration (PbS NCs) also increases the spectral sensitivity in the near-IR region. It is evident that this increase in responsivity of the device is due to increased electron transfer from the PbS NCs to C60. The peak in the near-IR region (~ 880nm) in the spectral responsivity corresponds to the first excitonic absorption of the nanocrystals used. Figure S6d shows the I-V characteristics of the higher PbS NC doping concentration device in the dark and under light (620nm) conditions. For completeness we also provide typical J-V characteristics of all photodoped devices obtained under dark conditions. It can be observed that upon doping the dark current is increased typically depending upon the type of donor material. We note that as device fabrication and photodoping process was undertaken by simple dropcasting procedures, dark currents in devices were found to be varying using the same photodopant materials.