Research Article www.acsami.org
Designing the Interface of Carbon Nanotube/Biomaterials for HighPerformance Ultra-Broadband Photodetection Youpin Gong,*,†,‡ Puja Adhikari,§ Qingfeng Liu,† Ti Wang,† Maogang Gong,† Wai-Lun Chan,† Wai-Yim Ching,§ and Judy Wu*,† †
Department of Physics and Astronomy, University of Kansas, Lawrence, Kansas 66045, United States College of Physics, Optoelectronics and Energy, Key Lab of Advanced Optical Manufacturing Technologies of Jiangsu Province, and Key Lab of Modern Optical Technologies of Education Ministry of China, Soochow University, Suzhou 215006, China § Department of Physics and Astronomy, University of MissouriKansas City, Kansas City, Missouri 64110, United States ‡
S Supporting Information *
ABSTRACT: Inorganic/biomolecule nanohybrids can combine superior electronic and optical properties of inorganic nanostructures and biomolecules for optoelectronics with performance far surpassing that achievable in conventional materials. The key toward a high-performance inorganic/ biomolecule nanohybrid is to design their interface based on the electronic structures of the constituents. A major challenge is the lack of knowledge of most biomolecules due to their complex structures and composition. Here, we first calculated the electronic structure and optical properties of one of the cytochrome c (Cyt c) macromolecules (PDB ID: 1HRC) using ab initio OLCAO method, which was followed by experimental confirmation using ultraviolet photoemission spectroscopy. For the first time, the highest occupied molecular orbital and lowest unoccupied molecular orbital energy levels of Cyt c, a well-known electron transport chain in biological systems, were obtained. On the basis of the result, pairing the Cyt c with semiconductor single-wall carbon nanotubes (s-SWCNT) was predicted to have a favorable band alignment and built-in electrical field for exciton dissociation and charge transfer across the s-SWCNT/Cyt c heterojunction interface. Excitingly, photodetectors based on the s-SWCNT/Cyt c heterojunction nanohybrids demonstrated extraordinary ultra-broadband (visible light to infrared) responsivity (46−188 A W−1) and figure-of-merit detectivity D* (1−6 × 1010 cm Hz1/2 W−1). Moreover, these devices can be fabricated on transparent flexible substrates by a low-lost nonvacuum method and are stable in air. These results suggest that the s-SWCNT/biomolecule nanohybrids may be promising for the development of CNT-based ultra-broadband photodetectors. KEYWORDS: s-SWCNTs, cytochrome c, ab initio calculation, exciton dissociation, broadband photodetectors
1. INTRODUCTION Photodetectors capable of detecting light in broadband underpins diversified optoelectronic applications in biomedical imaging, remote sensing, spectroscopy, and optical communication. Semiconducting single-wall carbon nanotubes (sSWCNTs) are a promising candidate material for ultrabroadband photodetectors. The key challenge to obtain highresponsivity and high-detectivity photonic carbon nanotube (CNT)-based photodetectors stems from the difficulties in dissociation of the photoexcited excitons. In single-walled carbon nanotubes (SWCNTs), the much enhanced Coulomb interaction and reduced dielectric screening as the excitons are confined within a small dimension of 1−2 nm comparable to the SWCNT diameter result in an unusually high binding energy on the order of few hundreds of meV.1−4 It is therefore critical to implement an effective exciton dissociation mechanism to separate the electron−hole pairs to form a photocurrent (Iph). Nanohybrids that interface inorganic nanostructures with compatible media can provide a unique © XXXX American Chemical Society
scheme for high-efficiency exciton dissociation and charge transfer through atomistic design and engineering of the interface heterojunctions. An optoelectronic nanohybrid has unique advantages over conventional materials because of (1) strong quantum effect in inorganic nanostructures such as SWCNTs, which yields superior electronic and optoelectronic properties such as higher light absorption, charge mobility, and spectra tunability; and (2) efficient exciton dissociation and charge transfer facilitated by the interface heterojunctions with appropriate interfacial electronic structures for photocarrier generation. A particular advantage of the carbon/biomolecule nanohybrid lies in its molecular-scale design of material building blocks that can have light-solid interactive properties superior to conventional materials and can be made in thin films for large-scale device fabrication with compatibility to Received: January 8, 2017 Accepted: March 6, 2017 Published: March 6, 2017 A
DOI: 10.1021/acsami.7b00352 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX
Research Article
ACS Applied Materials & Interfaces existing Si-based microfabrication procedures.5,6 While many works have been published in carbon/polymer composites,7−11 the success is limited so far in achieving high quantum efficiencies in optoelectronic nanohybrids.7,12,13 Biological macromolecules such as DNA and protein may present a promising alternative to polymers14 and have exhibited a great advantage in forming heterojunctions via either covalent or noncovalent bonds with carbon and other nanostructures.15−17 Besides the extraordinary and complex electronic structures that can hardly be replicated in synthetic materials, many biomolecules also have superior optical functionalities. The inorganic/biomolecule optoelectronic nanohybrids can therefore combine the unique electronic and optoelectronic properties through atomic-scale to device-scale design of photon absorption, exciton dissociation, charge carrier transfer, and transport. This provides tremendous opportunities for novel optoelectronic devices of extraordinary performance of high quantum efficiencies. In a recent work, we explored biomolecule cytochrome c (Cyt c, PDB ID: 1HRC) adsorbed onto s-SWCNT to form a molecular building block. The obtained high responsivity and high external quantum efficiency (EQE) up to ∼90% in nearinfrared (NIR) detection are attributed the high IR absorption of s-SWCNTs and the effective exciton dissociation at the sSWCNT/Cyt c interface.15 While the achieved EQE represents a more than two orders of magnitude enhancement over the previously reported on CNT-based nanocomposites,12 several fundamental questions remain. First, the exact band-edge alignment at the s-SWCNT/Cyt c interface and hence the efficiency of the exciton dissociation across the interface is unknown due to lack of knowledge on the band structure and density of states (DOS) in the Cyt c biomolecules. In fact, determination of these physical properties is by no means trivial for large molecules such as Cty c, having thousands of atoms. Without such a quantitative understanding, the applications of the s-SWCNT/Cyt c nanohybrids, or in general inorganic nanostructure/biomolecule nanohybrids, may be limited despite a broad spectrum of potential applications. On the other hand, Cyt c is a well-known optical molecule with absorption observed in the visible spectrum.18 If the sSWCNT/Cyt c nanohybrids are applied in the ultra-broadband from the visible region to NIR, both constituents s-SWCNT and Cyt c could be optically active and may contribute to the optoelectronic process. Experimental demonstration of this is important for the design of novel devices based on the nanohybrids but has not been done so far. Finally, the pairing of the s-SWCNT and Cyt c can potentially enable high photoconductive gain considering the strong quantum confinement in both SWCNT (diameter ∼ 1−2 nm) and Cyt c (diameter ∼ 3 nm) and the high charge mobility in SWCNTs. This can lead to significant enhancement of EQE much beyond 100%. In this letter, we carried out a detailed ab initio calculation of the electronic structure of one of the Cyt c macromolecules (PDB ID: 1HRC) using the OLCAO method.19 We obtained the HOMO (highest occupied molecular orbital) and LUMO (lowest unoccupied molecular orbital) based on the fully relaxed structure with high accuracy (see details in Figures S1 and S2, Supporting Information). Ultraviolet photoemission spectroscopy (UPS) was employed to experimentally measure the electronic structures of the PDB ID: 1HRC. This interactive approach allowed us to determine the band-edge alignment at the s-SWCNT/Cyt c heterojunction interface,
which is expected to provide a highly efficient exciton dissociation and charge transfer for ultra-broadband (visible light to NIR) photodetection. A remarkable EQE up to 5.9 × 104% has been achieved for the photonic photodetection derived from a considerable photoconductive gain up to 100− 500 in the s-SWCNT/Cyt c heterojunction photodetectors by minimizing the thermal loss. This leads to extraordinary responsivity (46−188 A W−1) from visible light to infrared and high figure-of-merit detectivity D*. In particular, the D* exceeding 1010 cm Hz1/2W−1 for uncooled NIR detection represents not only an enhancement of nearly 2 orders of magnitude over the best previous report on CNT-nanocomposites12,15 but also an improved performance of an order of magnitude over the current commercial VOx uncooled IR detector.20,21
2. MATERIALS AND METHODS Fabrication of s-SWCNT/Cyt c Films Photodetector. The detailed synthesis process has been described elsewhere.15 Briefly, the suspensions of Cyt c (from equine heart, purity ≥95%) and s-SWCNT (purity of semiconducting SWCNTs ∼ 95% with diameters of 1.2−1.7 nm and a length ranging from 300 nm to 5 μm) were mixed at 1:40 (sSWCNTs:Cyt c) mass ratio, and then the mixture was sonicated (Branson1800) in an ice water bath for 2−3 h. After the samples were centrifuged to remove insoluble material, the s-SWCNT/Cyt c films were fabricated using a vacuum filtration apparatus by employing a 0.2 μm filter membranes of mixed cellulose ester (MCE). The thickness of the formed films is ∼150 nm, close to the optimal SWCNT thickness (80−110 nm) for a large light absorption. Two Au(40 nm)/Ti(4 nm) metals as the source and drain electrodes with spacing of 0.3−0.4 mm were predeposited onto both flexible PET and SiO2(90 nm)/Si substrates by electron-beam evaporation through a shadow mask. The s-SWCNT/Cyt c film (width 0.2−0.4 mm) was transferred onto the target substrate with the Au/Ti electrodes; subsequently, it was placed in acetone to dissolve the filtration membrane. The optical absorbance spectra of the pure s-SWCNTs, the pure Cyt c (200 μg mL−1), and the s-SWCNT/Cyt (200 μg mL−1) solution were obtained by a ultraviolet−visible−NIR dual-beam spectrophotometer (Cary 5000, Varian). The microstructure of s-SWCNT/Cyt c nanohybrids was characterized by high-resolution electron microscopy (HRTEM) (Tecnai F20 XT, FEI). Photoresponse Measurements. The photoconductivity measurements were carried out under air atmospheric conditions at room temperature. The measured circuit was set up as voltage source mode in which a standard resistor (R0) was connected in series with the device, where the R0 value is approximately one percent of the device resistance. The voltage across the standard resistor in series was measured by a voltmeter (34420A, HP) when bias voltage was applied in the circuit using a voltage source (E3631A, Agilent), and then the current in the circuit was obtained by dividing measured voltage by the standard resistor. The illumination was provided by a 75 W xenon light with a 0.125 m gating monochromator (Cornerstone 130 1/8 m monochromator, Newport). The incident light power was monitored by a multifunction optical power meter (70310, Newport) and a UVenhanced silicon photodiode. The light density illuminated on the device surface depends on the wavelength range (300−1000 nm) and spot diameter. For time-resolved photoresponse measurement, 1000 and 532 nm light were modulated at various frequencies by a mechanical chopper (SR540, Stanford Research) and then illuminated onto the device; the time-resolved current output from the device was monitored by an oscilloscope (54624A, Agilent). The noise spectra were measured by a spectrum analyzer (SR760, Stanford Research) when various bias voltage was applied by a voltage source (E3631A, Agilent). Ultraviolet Photoemission Spectroscopy Measurements. A standard He discharge lamp is used as the excitation source. The photon energy used is 21.22 eV. The kinetic energy of the B
DOI: 10.1021/acsami.7b00352 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX
Research Article
ACS Applied Materials & Interfaces
Figure 1. (a) Calculated total DOS of 1HRC. (b) UPS spectrum of the Cyt c sample with the estimated HOMO level. The energy is referenced to the vacuum level (Ev). Photon energy of 21.22 eV was used for photoemission. The red line shows the UPS spectrum of Cyt c before the background subtraction. The black dashed line shows the secondary electron background fitted using a single exponential function. (c) HRTEM image of typical SWCNT absorbing a Cyt c. (d) Schematic of band alignment at the s-SWCNT/Cyt c interface for photoexcited carrier transport.
Figure 2. (a) Optical absorbance spectra of the pure s-SWCNTs (red), Cyt c (black), and s-SWCNT/Cyt c (blue) solutions. (b) Imaginary part of the complex dielectric function ε2 versus optical transition energy of 1HRC. Inset shows the local region of ε2 versus energy in panel b ranging from 0 to 4.5 eV. photoelectrons was detected by a hemispherical electron energy analyzer (Phoibos 100, SPECS). Electronic Structure Calculation. The electronic structure calculation of Cyt c (1HRC) was done based on the ab initio OLCAO method19 using a full basis set of atomic orbitals. The calculated results include the density of states, HOMO−LUMO gap and amino acid and heme group resolved partial charge distribution and interband optical absorptions (see the Supporting Information for more details).
aligning the two major peaks P1 and P2 with the calculated DOS and the extracted data from UPS together from the sharp conduction band (CB) edge to locate the LUMO (Figure S3, Supporting Information), we can estimate the band gap of 1HRC to be around 1.85 eV. It should be made clear that the term “gap” used here is loosely defined in terms of inorganic crystals which should be equivalent to the HOMO−LUMO gap in biomolecules. In the present context, the HOMO (−7.24 eV) and LUMO (−4.92 eV) levels for Cyt c are determined by a combination of experimental data and the ab initio calculations because the latter contains gap states that originate from heme group in 1HRC. On the basis of the electronic structure of Cyt c in Figures 1a and b, the heterojunction band-edge alignment at the interface of s-SWCNT/Cyt c nanohybrids, shown in the HRTEM images (Figure 1c and Figure S4 in the Supporting Information), can be determined using the HOMO (−5.44 eV) and LUMO (−4.56 eV) levels for SWCNT.22,23 Figure 1d depicts the interface heterojunction energy band offset of the sSWCNT/Cyt c nanohybrid. The conduction band (or LUMO)
3. RESULTS AND DISCUSSION Determination of Band-Edge Alignment at the sSWCNT/Cyt c Heterojunction Interface. The calculated total density of states for 1HRC is shown in Figure 1a. The spectrum is very complicated but, with careful inspection, it is possible to make a logical connection with experimental UPS measurements. The experimental value for the HOMO edge of Cyt c is at −7.24 eV relative to the vacuum level. The energy scale used for the UPS experiment is different from the one in the calculation (the zero of the energy is set at the highest occupied molecular orbital, or HOMO) (Figure 1b). By C
DOI: 10.1021/acsami.7b00352 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX
Research Article
ACS Applied Materials & Interfaces
Figure 3. (a) Schematic diagram of the optoelectronic process of the s-SWCNT/Cyt c nanohybrid photodetector. (b) Current versus bias voltage at various light wavelengths in the solar spectrum with the same order of magnitude optical power. Optical image of s-SWCNT/Cyt c photodetector on PET flexible (right inset) or SiO2/Si substrate (left inset). (c) Photocurrent of bias dependence under various excitation wavelengths from visible light to NIR with a comparable incident light power. Inset plots photocurrent as functions of incident light power at wavelengths of 400 nm (blue) and 1000 nm (red). (d) Representative dynamic photoresponse of the s-SWCNT/Cyt c nanohybrid devices at 15 V bias under 532 and 1000 nm wavelengths with modulation of 97 Hz.
the optical properties could be analyzed using ab initio wave functions with the full inclusion of the dipole transition matrix.19 The calculated optical spectrum is shown in Figure 2b in the form of imaginary part of the dielectric function. It can be seen that there are 2 prominent structures at around ∼0.67 and ∼2.5 eV (inset in Figure 2b), which could correspond to the same two prominent structures shown in the absorption curve of Cyt c at ∼550 and ∼408 nm, respectively.15 The origin of these transitions and a more accurate determination of the transition energy due to the complication from the presence of heme-related states within the nominal HOMO−LUMO gap is a subject of current investigation. However, it is quite clear that ab initio calculations in Cyt c enable us to predict optical absorptions of the bulk Cyt c in the inorganic/Cyt c nanohybrids in further improving the device performance by entangling and fine-tuning the various attributes related to the device performance. The broad peak centered around 13−15 eV is characteristic of optical transitions for many similar biomolecular systems.19,28−31 Optoelectronic Process and Properties. Figure 3a illustrates the device structure of a photoconductive photodetector made using s-SWCNTs/Cyt c nanohybrids for investigation of the optoelectronic properties. While sSWCNTs contribute dominantly to the light absorption in the solar spectrum, Cyt c is expected to enhance the absorption at shorter wavelengths
