In-situ Synthesis of Carbon Nanotube-Graphite. Electronic Devices and Their Integrations onto. Surfaces of Live Insects and Plants

Transcription

1 Supporting Information In-situ Synthesis of Carbon Nanotube-Graphite Electronic Devices and Their Integrations onto Surfaces of Live Insects and Plants Kyongsoo Lee,, Jihun Park,, Mi-Sun Lee,, Joohee Kim,, Byung Gwan Hyun,, Dong Jun Kang, Kyungmin Na, Chang Young Lee, Franklin Bien, Jang-Ung Park,,* School of Materials Science and Engineering, Low-Dimensional Carbon Materials Research Center, Ulsan National Institute of Science and Technology (UNIST), Ulsan Metropolitan City, , Republic of Korea Nano-Convergence Devices Research Group, School of Nano-Bioscience and Chemical Engineering, Ulsan National Institute of Science and Technology (UNIST), Ulsan Metropolitan City, , Republic of Korea Creative and Fundamental Research Division, Korea Electrotechnology Research Institute, Changwon, , Republic of Korea School of Electrical and Computer Engineering, Ulsan National Institute of Science and Technology (UNIST), Ulsan Metropolitan City, , Republic of Korea This PDF file includes: Methods Supporting Figures S1-S6 Supporting Table S1 Caption of Supporting Movie

2 Methods In-situ synthesis of SWCNTs-graphite. Catalysts patterns for the in-situ synthesis of graphite and SWCNTs were prepared using a lift-off process. A 300 nm-thick SiO 2 was used as the gate dielectric, and it was deposited using sputter on a Cu foil (thickness: 50 μm). Lift off resist (LOR 3A, MicroChem) and positive photoresist (s1818, MicroChem) were spin-coated and, then used to pattern the electrodes on the SiO 2 /Cu foil. Subsequently, a 150 nm-thick Cu layer and a 150 nm-thick Co layer, catalyst for graphitic electrode, were deposited sequentially by thermal evaporation without breaking a vacuum, and formed the catalyst patterns successively after the remaining lift off resist and photoresist were washed out. Then, thermal evaporation of a very thin Fe (thickness: 6 Å) also was deposited on the surface of the SiO 2 for the synthesis of random network SWCNTs after channel pattern with thin photoresist (S1805, MicroChem), followed by removal of the photoresist by a solvent. The catalysts pattern for graphite with horizontally aligned arrays of SWCNTs, instead of the random network forms of SWCNTs, were prepared by replacing the SiO 2 /Cu and Fe patterns by a stable temperature (ST)-cut quartz wafer and Fe pattern with a narrow line at the end of the electrode, respectively. These patterns were pre-annealed at 925 C for 2 hours prior to carbon growth, to form metal alloys by local diffusion and Fe nanoparticles by dewetting on the SiO 2 surface. After loading the samples onto the center of a quartz CVD chamber at atmospheric pressure, the furnace temperature was increased to 950 C with Ar (300 sccm) and H 2 flows of 300 sccm each. In-situ synthesis was conducted under the flow of CH 4 (500 sccm) and H 2 (75 sccm) for 5 minutes, after which the chamber was cooled to room temperature. Fabrication of in-situ synthesized SWCNTs-graphite back gate FET and functionalized gas sensor. A spin-coated thick SU8 photoresist (thickness: 1 μm) was patterned onto the in-situ synthesized SWCNTs-graphite structure with the openings around the SWCNT channel. Then a 200 nm-thick poly (methyl methacrylate) (MicroChem Corp) supporting material was spin coated on the synthesized SWCNTs-graphite structure. The Cu, Co and Fe metal catalysts were dissolved in a

3 diluted etching solution of FeCl 3 :HCl:H 2 O (1:1:20 vol.%), and a SiO 2 layer and quartz wafer also were dissolved in buffered oxide etchant, HF:H 2 O (6:1 vol.%) with the PMMA-coated SWCNTsgraphite structure floating on the surface of the solution. Subsequently the samples were transferred onto a target substrate. Specifically, the samples were transferred onto a Si wafer with a 300 nm-thick thermal oxide layer to evaluate the electrical properties of the SWCNTs-graphite FET synthesized in situ. Then, the PMMA supporting material was removed with diluted acetone. The samples were drop coated with 5 wt.% PPy aqueous solution (Sigma Aldrich) for 5 minutes and rinsed with DI water for functionalization of the SWCNTs for gas sensing. Optical characterization. The Raman spectra were recorded with a WITec CRM200 Raman system with a 532 nm laser as the excitation source. The optical transmittance of films was measured using UV-vis-NIR spectroscopy (Cary 5000 UV-Vis-NIR, Agilent). The transmittance of the substrates was excluded. TEM characterization. The structure of the SWCNTs-graphite was transferred onto a micromachined holey gold grid (4220G-XA (R 1.2/1.3), SPI Supplies). High-resolution transmission electron microscopy (HRTEM) and diffraction pattern analyses were conducted in a JEOL, JEM The cross-sectional image was observed at an acceleration voltage of 200 kv by focusing the edge area of the SWCNTs-graphite interface. Electrical characterization. Electrical characterizations were conducted with the I d -V g, I d -V d and R-T characteristic measurement using a Keithley 4200-SCS semiconductor parametric analyzer. R-T characteristics upon exposure to DMMP vapors were monitored at V d = 1 V and V g = 0 V. Insect anesthesia. To handle the insect without damaging them, the insect was anesthetized with diethyl ether. To administer the anesthesia, a stag beetle was placed alone in a plastic bottle (30 cm wide 10 cm long 10 cm high) with cleanroom paper that was wet by 2 ml of diethyl ether in a vial.

4 The insect was observed until it stopped moving (generally for 30 minutes) in order to immediately remove the vial from the plastic box. Most of the insects completely recovered from the anesthesia after transfer and measurement. DMMP detection measurements. A PPy aqueous solution (5 wt.%, Aldrich) was dropped onto the SWCNTs network and rinsed after 10 minutes. Resistance changes in the sensor upon exposure to DMMP vapor were monitored at V d = 1 V, V g = 0 V (Keithley 4200 SCS). The DMMP vapor was injected through an aligned syringe to the PPy coated SWCNTs network using a syringe pump at room temperature and ambient conditions. The initial concentrations of DMMP were diluted with air to attain the concentrations. After the device as attached to a live insect, measurements were taken after the anesthetized insect fully recovered from the anesthesia (Supporting Movie S1). Wireless sensing measurement. The wireless sensor was designed to have an LRC passive circuit with a resonant frequency of ~400 MHz using Ansoft HFSS software. The wireless sensor consisted of two resonant coils. The outer coil was coupled inductively to an inner coil that was coupled inductively to a resistive sensor. The planar square helix coils were made of an electrically conducting Au wire (thickness: ~150 nm) with a total length of ~50 cm wound into a helix of seven turns in an outer coil and a total length ~5.6 cm with a single turn helix in an inner coil. The wireless sensor was transferred onto a stag beetle in the manner described above. The wireless sensor was tested by a network analyzer (Agilent, N5227A) at DMMP concentrations of 5, 10, and 15 ppm. A reader antenna was used to inductively couple and power the remote sensor. The relative differences in resistance of the SWCNTs-graphite sensor were used to derive the changes in the reflection value (S11 parameter) characteristic at the sensor resonance frequency.

5 Supporting Figures Figure S1. AFM images of Fe nanoparticles. (a) AFM image of the thermally deposited 6 Å Fe nanoparticles (top) and their height profile along the dash (bottom). (b) AFM image of the thermally deposited 3 Å Fe nanoparticles (top) and their height profile along the dash line (bottom). The results show that density of Fe nanoparticles depends on the thickness of the deposited film. Scale bars, 2 nm.

6 Figure S2. HR-TEM images of SWCNTs-graphite contact areas. HR-TEM image of the contact between SWCNTs and graphite shows that SWCNT and graphite do not form the covalent sp 2 -sp 2 bonding. Scale bars, 5 nm. Electron diffraction pattern, which taken from graphite, indicates the polycrystalline nature. Scale bar, 5nm -1.

7 Figure S3. Optical transmittance spectra of SWCNTs network and graphite. SWCNTs with the random network exhibit relatively high transmittances above 95% at 550 nm in wavelength (left of the inset). And the transmittance of the synthesized graphite (thickness: 150 μm) is ~15% at 550 nm (right of the inset). Scale bar, 1 cm.

8 Figure S4. Effect of fitted device to a leaf of live plant. (a) A plant with freshly-attached device. (b) The same plant after one month with the device still attached. The plant remains healthy with no noticeable sign of discoloration or withering of the leaf. Scale bars, 1 cm.

9 Table S1. Repeatability and reproducibility of the sensors on three different types of surfaces (SiO 2, leaf and insect). Reproducibility of 108 devices in total was investigated with the DMMP exposure at concentration of 5 ppm. Also, the repeatability was examined by exposing each device to the DMMP vapor (5 ppm) for 3 seconds with intervals over 10 minutes repeatedly (the number of repetition: 50 times).

10 Figure S5. Real-time sensing characteristics of the sensor on the SiO 2 surface, by exposing the sensor to DMMP in air with different humidity (left graph: 25 % RH, and right graph: 70 % RH).

11 Figure S6. Wireless monitoring of DMMP vapor. (a) A photograph of the wireless sensor integrated onto the epidermis of the insect, Scale bar, 1 cm. (b) The reflection (S11) changes of the wireless sensor at varied DMMP concentrations. The reflection (S11) at the resonance frequency decreased with increase in the DMMP concentrations.

12 Calculation of mobility by rigorous model Parallel plate model is commonly used to calculate the mobility as performance of common FETs with a sheet channel. In the case of FET based on carbon nanotube, the parallel plate model is not suitable to apply for the gate capacitance ( ) when the density of nanotubes ( ) is low. A more accurate model that is considered the realistic electrostatic coupling between sparse carbon nanotubes and the gate electrode can be used to estimate the gate capacitance of carbon nanotube based FET. The carrier mobility ( ) in parallel plate model is evaluated by = Effective field-effect mobility by accurate capacitance model of the carbon nanotube based FET as = where is the capacitance per unit area. The total charge density of the FET with density of carbon nanotubes per unit width, single carbon nanotube capacitance calculated by = = log 2 where is the quantum capacitance of carbon nanotube, and is the radius of carbon nanotubes. We evaluated the accurate mobility with the above expression. In this equation, the average radius is 2 nm as shown Figure 1c, and the density of SWCNTs was ~1 tube/μm, as determined by Figure 1a.

13 Wireless sensing measurement. The reflection at the resonance frequency is inverse proportion to the electrical resistance in the SWCNTs channel as resistive element of the circuit. The expression for which is, using the Kirchhoff's circuit laws, = 2 + where, is the resistance of the sensor, is the mutual inductance between inner coil and outer coil, is also the mutual inductance between outer coil and read coil and is the resistance of the reader, respectively. The above expression is expressed in liner form. When the S parameter magnitude is expressed in logarithmic form (unit of decibels), magnitude is agreement with our result.

14 Captions of Supporting Movie Supporting Movie S1. For integration with a live insect, transfer and measurement were carried out after anesthesia. Insect fully recovered from anesthesia after transfer and measurement.