Batch Fabrication of Ultrasensitive Carbon Nanotube Hydrogen

Subscriber access provided by UNIV OF DURHAM

Batch Fabrication of Ultrasensitive Carbon Nanotube Hydrogen Sensors with sub-ppm Detection Limit Mengmeng Xiao, Shibo Liang, Jie Han, Donglai Zhong, Jingxia Liu, Zhiyong Zhang, and Lian-Mao Peng ACS Sens., Just Accepted Manuscript • Publication Date (Web): 05 Apr 2018 Downloaded from http://pubs.acs.org on April 5, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Sensors

Batch Fabrication of Ultrasensitive Carbon Nanotube Hydrogen Sensors with sub-ppm Detection Limit Mengmeng Xiao†, Shibo Liang†, Jie Han, Donglai Zhong, Jingxia Liu, Zhiyong Zhang* , Lianmao Peng*

Key Laboratory for the Physics and Chemistry of Nanodevices and Department of Electronics, Peking University, Beijing 100871, China *Correspondence to: E-mail: (Z. Y. Z.) [email protected]; (L. M. P.) [email protected]. †These authors contributed equally to this work ABSTRACT Carbon nanotube (CNT) has been considered as an ideal channel material for building highly sensitive gas sensors. However, the reported H2 sensors based on CNT always suffered from the low sensitivity or low production. We developed the technology to massively fabricate ultra-highly sensitive H2 sensors based on solution derived CNT network through comprehensively optimizing the CNT material, device structure and fabrication process. In the H2 sensors, high semiconducting purity solution-derived CNT film sorted by PFO polymer is used as main channel, which is decorated with Pd nanoparticles as functionalization for capturing H2. Meanwhile Ti contacts are used to form Schottky barrier for enhancing transferred charge-induced resistance change, and then a response of resistance change by three orders of magnitudes is achieved at room temperature under the concentration of ~311 ppm with a very fast response time of approximately 7 s and a detection limit of 890 ppb, which is the highest response to date for CNT H2 sensors and the very first time to show the sup-ppm detection for H2 at room temperature. Furthermore, the detection limit concentration can be improved to 89 ppb at 100 C°. The batch fabrication of CNT film H2 sensors with ultra-high sensitivity and high uniformity is ready to promote CNT devices to application at first in some special field.

1 / 21

ACS Paragon Plus Environment

ACS Sensors 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 21

KEYWORDS: Carbon nanotube, hydrogen sensor, network film, high sensitivity, detection limit There is a continued need for faster, more accurate and more selective detection of hydrogen gas (H2) in connection with the development and expanded use of hydrogen gas as an energy carrier, as a chemical reactant and as a biomarker of some disease.1-3 Amongst all kinds of H2 sensors, resistor-type sensor based on semiconductors presents lower cost, smaller size and easier integration, and then get more attention in the future applications.4-7 Single-walled carbon nanotube (SWCNT) is an ideal material for building high sensitivity chemical sensors mainly owing to its ultra-high surface to volume ratio, which means each atom is physically accessible under ambient conditions.8 Since the first demonstration of CNT as a sensor for detecting H2 by Kong et al,9 a number of H2 sensors have been manufactured based on individual CNTs or CNT films.10-15 Although individual semiconducting CNT is the excellent channel to realize high-sensitivity H2 sensors, this kind of CNT material cannot support massive preparation of any device, which limits the wide applications of CNT sensors. H2 sensor based on CNT film tends to be fabricated in batch through a conventional micro-manufacture process, but is easy to suffer from low sensitivity owing to low performance and low semiconducting purity of CNTs. Generally, the metallic CNTs remains in the film blocked the pursuing on high sensitivity in CNT film based gas sensors to date10, 11, 15-18 and then the best detection limit of CNT film based H2 sensors keeps at only 10 ppm at room temperature.15 Although the best detection limit as high as 1 ppm at room temperature was achieved in CNT film based H2 sensors, the corresponding response was very small.19

In recent years, significant

progresses have been achieved on solution processed CNT films, in which the semiconducting purity has been improved up to 99.99%.20, 21 It is just the time to develop compact CNT based H2 sensors featuring high sensitivity and selectivity, as well as batch fabrication and low cost. Especially in some special application, such as monitoring H2O leakage to the sodium in fast nuclear reactors,22 or hydrogen breath 2 / 21


Page 3 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Sensors

tests,1 ultra-highly sensitive compact H2 sensors with the detection limit lower than 1 ppm and high response are eagerly required, and CNT film based H2 sensor can be promote to reach this level. In this report, we massively fabricated ultra-highly sensitive H2 sensors based on solution derived CNT network through comprehensively optimizing the CNT material, device structure and fabrication process. In the H2 sensors, we used the high semiconducting purity solution-derived CNT film sorted by PFO polymer as main channel, and decorated the channel with Pd nanoparticles as functionalization for capturing H2. Meanwhile Ti contacts are used to form Schottky barrier for enhancing the effect of charge transfer on resistance change of the sensor, and then a response of resistance change by three orders of magnitudes is achieved in the sat room temperature under the concentration of ~311 ppm. The fabricated CNT based H2 sensors present a very fast response time of 7 s at 311ppm and a detection limit of 890 ppb, which is the highest response to date for resistor-based sensors and the first time to show the sup-ppm detection for H2 at room temperature. Furthermore, the detection limit concentration can be improved to 89 ppb (limited by the measurement setup) at 100 ℃. The massive fabrication of CNT film H2 sensors with ultra-high sensitivity and high uniformity is ready to promote CNT devices to application at first in some special field. RESULT AND DISCUSSION As is well known, the electrical properties of semiconducting CNTs are very sensitive to transferred charges from gas molecule, while metallic ones are inert to molecule absorption which does not significantly change the density of states near the Fermi level8. Therefore, the high-quality CNTs film with high semiconducting purity is the basic material for constructing ultra-highly sensitive H2 sensor. To explore the effect of semiconducting purity in CNT film on performance of H2 sensor, two kinds of CNTs, i.e. the raw one (or r-CNT, for short) with natural semiconducting purity and sorted one (or s-CNT, for short) with high semiconducting purity, are used in this 3 / 21


ACS Sensors 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 21

work. Figure 1a and b respectively show the UV-Vis-NIR absorption and Raman spectroscopy of the two kinds of CNT solutions. Compared with r-CNT film, the invisible ﬁrst metallic excitonic transitions peak (M11 between 600-800 nm in Figure 1a) and sharp S22 peak (820-1350 nm) for s-CNT indicate an ultrahigh semiconducting purity. Further characterization of short channel devices shows that the semiconducting purity higher than 99.9%.23, 24 The narrow RBM peak at 160 cm-1 in sorted CNT solution (Figure 1b) indicates a narrow diameter distribution around 1.5 nm according to the experimental relationship of ω=248/d (nm). High IG/ID ratio suggests few defect in CNTs used here. The SEM image (Figure 1c) of the sorted CNT film shows an average length of approximately 2 µm and a density of approximately 20 tubes/µm with good uniformity in large scale. Furthermore, the as-fabricated CNT film is a monolayer one since it exhibits a thickness of 2-3 nm measured by AFM (Figure 1d). It is worthy to mention that monolayer film with suitable density is important for sensing, because high density will cause screening effect between CNTs thus lowering the sensitivity.

25

Therefore, the high quality and

ultra-high semiconducting monolayer CNT film with suitable density and good uniformity has been prepared for constructing high performance H2 sensors in batch. Taking advantage of the full surface coverage of CNT network on substrate, we can massively fabricate resistor-type H2 sensors through a photolithography-based process, and the as fabricated devices are shown in Figure 2a, which is consist of 24 sensors as one unit. The structural diagram of an individual H2 sensor is shown in Figure 2b, which is similar to conventional back-gated FET based on CNT network film. Three points in the H2 sensors are well designed differing from conventional FET. At first, the device is designed as a long channel one to highlight the channel resistance. Here an optimized channel length (Lc) of 10 µm was used to achieve the highest response as shown in Fig.S1. The channel was designed as wide as 100 µm to assure at least 2000 CNTs in one device, which guarantees the electronic uniformity of each sensor. Secondly, the CNT network channel is decorated with Pd nanoparticle with a nominal thickness of 1nm as the functional material for sensing H2. Thirdly, Ti film instead of 4 / 21


Page 5 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Sensors

Pd is used to form Schottky contact to CNT film, and meanwhile to avoid any H2 induced work function change on electrodes. Since intrinsic CNT is not sensitive to H2, Pd particles decoration in channel is necessary to act as efficient catalysts for adsorbing and decomposing H2. The thickness of Pd decoration in channel is a key parameter for our CNT based H2 sensors. The electron-beam evaporation deposited Pd should form discontinuous nanoparticles other than continuous film, as shown in the SEM picture in figure 2b and AFM picture in Figure S2a, ensuring that the electrical conduction route is only through the CNT film. On one hand, too thin Pd film decorated should lead to less sensitivity and worse uniformity on H2 sensors due to the less effective s-CNT and Pd/H atom interaction. On the other hand, with thicker Pd film, which is becoming continuous, metallic Pd film begins to dominate the conductance of channel. The metallic channel is not as sensitive to the transfer charges as semiconducting channel. Through the contrast experiments shown in Figure S3, Pd film with thickness of 1 nm is the optimized functional layer to achieve the high sensitivity for our H2 sensors, and then all of the following sensors are decorated by 1 nm Pd film. To check the effect of Pd decoration, the CNT film based sensors devices are measured as back-gated field-effect transistors (FETs), and the transfer characteristics before and after Pd nanoparticles decorating are shown in Figure 2c. The FETs based on sorted CNT film present 2 orders of magnitudes higher on/off current ratio than these on unsorted CNT film (see Figure S4), indicating higher semiconducting purity in CNT film after sorted. Obvious loss of gate dependence and decrease of on current are observed in both kinds of CNT film FETs after depositing Pd nanoparticles. Pd nanoparticles in CNT channel will screen gate electric field, and then channel current cannot be fully turned off in the same gate voltage, resulting in the on/off ratio reduction.26 Meanwhile, Pd decoration also induces charge transfer between CNT and Pd, resulting in a localized depletion area surrounding the metal nanoparticle and acting as a scatter center for holes transport.27 The Pd-decoration induced current reduction in sorted CNT based FETs (47.6%) is much larger than that (16.6%) in 5 / 21


ACS Sensors 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 21

unsorted-CNT FETs since the semiconducting CNTs are more sensitive to charge transfer than the metallic ones. The H2 sensing performance of Pd nanoparticles decorated CNT film sensors was evaluated in N2 at ambient pressure. All the sensing measurements in this work were conducted at Vds=-0.1 V and Vgs-backgate =0 V at ambient temperature if not specified. Response is one of the most important parameters for a gas sensor, and is defined as

Response = ( R − R0 ) / R0 ,

(1)

where R and R0 are the channel resistances before and after sensing gas. As a result, as shown in the measurement results for H2 sensing in Figure 2d, the sorted CNT film based sensor presents much higher response (~1000) to 311 ppm H2 than the unsorted CNT film device by three orders of magnitudes. Contrast test executed on 1 nm Pd film (red curve in Figure 2d) indicates the high response is not originated from nanoparticles. It is noteworthy that 1000 is the highest response value at the similar concentration for H2 detection based on carbon nanomaterials. The transfer curves with/without H2 in s-CNT based sensors are shown in Figure 2e, which indicates H2 (with concentration of 311 ppm) leads to a current decrease by approximately three orders of magnitudes at any gate voltage. In comparison, the sensor based on r-CNT film exhibits a slight current decrease to 311 ppm H2 at any gate voltage (see the right panel of Figure 2e). The transient responses of the s-CNT sensors under different H2 concentrations are measured and shown in Figure 3a, indicating an increasing response with the concentration ranged from 0.89 ppm to 311 ppm (calculated by the mass controller flow rate) at room temperature. It should be noted that the sensor using s-CNT film exhibits the ability of detecting H2 concentration down to 890 ppb at room temperature (see the inset of Figure 3a), and the response (after 200 s H2 exposure) reaches up to approximately 50% at such a low concentration. It’s the first time to realize the sub ppm H2 detection through resistor-type sensors at room temperature. 2 The response increases with the H2 concentration as shown in figure 3b, which indicates that our sensors exhibits liner response at higher concentration range (40 6 / 21


Page 7 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Sensors

ppm-311 ppm) and deviates from the liner response at low concertation range. To study the selectivity of our sensors, we measured the sensor’s response to various gases, including 1000 ppm CH4,1000 ppm CO2,500 ppm NO2, 500 ppm H2S,pure O2, 3% H2O and 1% ethanol. As shown in the inset of Figure 3b, the H2 sensor exhibits invisible response to 1000 ppm CO2 (-2%) or 1000 ppm CH4 (-1.8%). Dynamic responses to other gases are shown in figure S5, indicating a high selectivity after decorating Pd. In addition, the transient response of two H2 sensors are measured at the same time for three cycles under the same H2 concentration of 311 ppm (Figure 3c), and the two devices exhibit almost identical transient response curves with a conductance change by three orders of magnitudes. By fitting portions of the curves to exponentials of the form e-t/τ in figure 3c, response time τresponse and decay time τdecay12 are 7 s and 89 s respectively. We further studied the reproducibility and stability of three more sensors by repeating the detection process for 15 cycles and measure them again after four days stored in ambient air (Figure S6). No obvious degradation of the baseline current and response was observed during the measurements, revealing excellent reproducibility, stability and recovery of our devices. Statistic gas sensing performances of 28 devices under different H2 concentration were shown in Figure S7 to show the batch fabrication of the high performance sensors. The reaction taking place during the H2 test is as9 H 2 → 2H atom-on-surface → H atom-in-Pd

O 2 + 2H atom-surface → 2OH,OH → H 2O .

(2) (3)

Before H2 exposure, the stabilized period is to exhaust the air in the test chamber and gas line to avoid its effect on sensor performance. 31 In the recovery period, the H atom should diffusion to the outer surface of Pd from bulk to react with the O2. This process will take a long time which is decided by the diffusion speed of H atom in Pd. We compared the response to H2 and other key metrics in our sensors with the previous reported CNT-based H2 sensors as shown in Figure 3d and Table 1. Our H2 sensors presented the best values on detection limit, response, and response time 7 / 21


ACS Sensors 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

simultaneously, and can be massively fabricated. Especially, the response of our s-CNT film sensors is 2-3 orders of magnitude higher than that of the previous reported CNT film based H2 sensors, and even is comparable with that of the most sensitive H2 sensors based on individual CNT.13 Furthermore, the s-CNT film sensors here present the much higher sensitivity and lower detecting limit than any reported CNT sensors, indicating our s-CNT film based H2 sensors are the ultra-sensitive one with ability to detect sub-ppm H2 at room temperature. It is the first time to achieve a response exceeding 100 on CNT network H2 sensor at any concentration and any temperature. The high response (and high sensitivity) of our H2 sensors are mainly originated from improvement on material, device design and fabrication. In the following, we will discuss the main reasons for obtaining large response in our sensors through exploring the H2 detecting mechanism as described in Figure 4a. At first, the highly semiconducting enriched CNT film is of great importance for the high sensitive sensors since semiconducting CNTs are more sensitive than the metallic ones. As a contrast, the r-CNT film based sensors present similar response to the reported CNT film sensors (see Figure 3d), indicating the main reason of our ultra-sensitive sensors is originated from the high semiconducting purity. Secondly, decoration of Pd nanoparticles with suitable thickness is crucial for the high sensitivity H2 detection in the CNT film sensors. Pd particles can act as an efficient catalytic for decomposing and adsorbing H2 at CNT film channel. H2 interacts with Pd nanoparticles through two ways as depicted in Figure 4b. On one hand, hydrogen dissolves into Pd to lower the work function of Pd. On the other hand, H2 dissociates on Pd to initiate the spillover of H atoms.28, 29 The former effects will lead to electron transfer from Pd to CNTs, which induces localized depletion area and increase of the resistance (denoted as R PdH-CNT in Figure 4a). The spillover H atoms will diffuse on CNT surface and also directly donates electrons to the CNT, inducing a delocalized depletion area and increase of the resistance (denoted as RCNT-H in Figure 4a).30 We named the comprehensive effect of these two effects as s-CNT and Pd/H atom interaction. As many enough depletion areas are overlapped to format a complete 8 / 21


Page 8 of 21

Page 9 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Sensors

barrier across the channel, the device is completely turned off, leading to large resistance change by several orders of magnitude. Thirdly, Ti film is used to form Schottky contact to CNT film in the sensors. On one hand, H2 induced the electrode work function change can be avoided since Ti is a H2 inert metal. What’s more, the Schottky barrier (SB) at Ti/CNT junction is in favor of achieving large response in the sensors as shown in Figure S8, which shows that the Ti –contacted sensor presents a ten times higher response than the Pd-contacted one. In an ohmic contact sensor, the H2 induced doping in the CNT channel will lead to linear resistance change on current. While for SB contact sensors, the H2 induced charge transfer will dope the CNT channel, and then tune the thickness of SB, which will bring exponential change on current. The detailed H detecting mechanism is described by the energy band diagram in Figure 4c. Before H2 in, the CNT network can be considered as a degenerate semiconductor due to heavily doped by Pd nanoparticles,26 and then hole can tunnel through the ultra-thin SB. As H2 was introduced, the thickness of SB increases with the decrease of doping concentration, and then tunneling current exponentially decreases. The Pd decoration captures and senses the H2 gas, and Ti contact amplifies the response. Therefore, under a low H2 concentration, the change of SB resistance mainly contributes to the whole resistance change, resulting in the high sensitivity and ultra-low detection limit in the H2 sensors. As the H2 concentration increases to a high level, thermal emission of hole becomes the main mechanism and the current begins to depend on the SB height which is constant, and then contact resistance no longer varies with H2 concentration. It means the response should be saturated at high H2 concentration or at a long enough period exposure to H2. In addition, the dissociation speed of H2 on Pd surface and the diffusion speed and length of H atom in Pd and on CNT surface will effectively influence the sensitivity. This diffusion process is expected to be influenced by the operating temperature.28 It is confirmed by measuring the response of the CNT film sensors at higher temperature environment, where the diffusion speed of H atom is improved. It is 9 / 21


ACS Sensors 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

obvious from Figure 4d that a 10-times improvement on response is achievable by increasing temperature from room temperature to ~100 ℃. As a result, the detection limit is improved to 89 ppb, which is among the best detectivity for H2 sensors.2 Besides, the effect of humidity was also studied (Figure S9). Water may dissolve on Pd surface, and the resulted O2- formation occupies the active site on Pd surface to reduce the response.10 This may be eliminated by introducing the nanofiltration on the channel material surface.31 CONCLUSIONS We have massively fabricated ultra-highly sensitive H2 sensors based on solution derived CNT network through comprehensively optimizing the CNT material, device structure and fabrication process. In the H2 sensors, high semiconducting purity solution-derived CNT film sorted by PFO polymer is used as main channel, which is decorated with Pd nanoparticles as functionalization for capturing H2. Meanwhile Ti contacts are used to form Schottky barrier for enhancing transferred charge-induced resistance change, and then a response of resistance change by three orders of magnitudes is achieved in the sat room temperature under the concentration of ~311 ppm. The fabricated CNT based H2 sensor presents a very fast response time of 7 s at 311ppm and a detection limit of 890 ppb, which is the highest response to date for CNT H2 sensors and the very first time to show the sup-ppm detection for H2 at room temperature. Furthermore, the detection limit concentration can be improved to 89 ppb (limited by the measurement setup) at 100 C°. The massive fabrication of CNT film H2 sensors with ultra-high sensitivity and high uniformity is ready to promote CNT devices to application at first in some special field. MATERIALS AND METHODS Preparation of CNT solution Raw Arc-discharged SWNTs were purchased from Carbon Solution Inc. The dispersants (poly[9-(1-octylonoyl)-9H-carbazole-2,7-diyl(PCz)]) were synthesized by 10 / 21


Page 10 of 21

Page 11 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Sensors

Suzuki poly condensation. 100 mg AP-SWCNTs and 100 mg PCz were added in 100 mL toluene. Then the solution was dispersed with a top-tip dispergator (Sonics VC700) at 300 W for 30 min. The dispersed solution was centrifuged for 0.5 h at 50000 g (Allegra X22R centrifuge) to remove most metallic CNTs and insoluble materials. The upper 90% of the supernatant was collected and centrifuged for additional 2 h at 50000 g. The as-collected upper 90% supernatant was used to prepare the CNT thin film with dip-coating method. Unsorted tube solution was obtained in the similar way except for the tubes were dispersed in chloroform. Preparation of CNT Thin Film and Fabrication of Gas Sensors The as-prepared SWNT solution was diluted with toluene or chloroform to obtain a solution of 50 ug/ml by absorption spectra. To prepare CNT films on substrate, the target silicon substrates with 500 nm SiO2 were submerged in the diluted solution for 48 hours. The substrate was then taken out of the solution, rinsed with toluene to remove the loosely bonded carbon nanotubes and excess sorted polymer. Then the sample was blew with 99.999% N2 and baked at 150 °C on hotplate for 30 min at atmosphere. The as-prepared sample was carefully annealed under H2/Ar mixture for 1 hour at 600 ℃ to remove partially of the sorted polymer bonded on the CNT surface (see Figure S10). Then the source and drain were defined by photolithography and Ti/Au (10 nm/50 nm) or Ti/Pd（0.3 nm/60 nm）was deposited by electron beam evaporation electrode, with a channel length of 10um and channel with 100 um. Pd film with different thicknesses was deposited over the whole channel using electron beam evaporation with a photolithography process to define the Pd area. Original AFM picture show a carbon nanotube density of 20-30/um with an average length of ~2 um. Raman spectrum and short channel characterization has been performed showing a nearly >99.9% semiconducting ratio. Measurement of H2 sensing performance The as-prepared device was wire bonded to a self-made printed circuit board (PCB) board and sealed in a home-made quarts chamber that is connected to a gas line 11 / 21


ACS Sensors 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

system and electrically connected to Kethily 4200 semiconductor analyzer. To reduce the ‘parasitic time’, we use a small quartz chamber with a volume of 375 ml and the connecting tube is as shorter as possible. It is reported that O2 and H2O adsorbates on Pd nanoparticles will decrease the sensor’s response to H2.31 So all of our tests start with an activation process to remove the ‘deep absorption’ of O2 and H2O on Pd nanoparticles. Briefly, the sensor array was exposure to the high concentration H2 (1%) flow for 10 min to reach a steady state then using synthetic air (O2/N2,20%/80%) to recover the sensor to the original state. After then, typical concentration test was performed. Total flow rate was maintained at 2000 sccm and the flow rates of individual gases were varied by mass flow controllers. H2 with different concentration was obtained by dilute the original 3% H2/N2 mixture with pure N2 gas. Each circle gas exposure consisted of a five-minute stabilized period with pure N2, 200 s exposure to target concentration of H2 gas, and another stabilized period in pure N2 for 100 s to see if the sensor will recover in N2, and then a recovery period in synthetic air. The recovery period is prolonged as H2 concentration increasing. A small bias of Vds=-0.1 V or -1 V was maintained to avoid the joule heating effect. At each concentration, sensor response increases as time prolonging with a different increasing rate. With a fixed exposure time of 200 s, the sensor response increases linearly with the gas concentration. As limited by our measurement systems, a saturation response is not visible till the H2 concentration increase to 311 ppm. All the sensing experiments were performed in atmosphere pressure and Vds=-0.1 V and Vgs-backgate=0 V. For the high temperature measurements, a heater resistor was attached to the PCB board.

Supporting Information. The AFM picture of Pd nanoparticles with different thickness and sensor performance, back-gate devices comparison of r- and s- CNT, selectivity of the Pd decorated s-CNTs film sensors, statistic gas sensing performances, stability, contact effect, channel dimension, material annealing and humidity effect on the sensor performance are available in the supporting information. 12 / 21


Page 12 of 21

Page 13 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Sensors

AUTHOR INFORMATION Corresponding Author *E-mail: (Z.Y.Z.) [email protected]. *E-mail: (L.M.P.) [email protected]. Notes The authors declare no competing financial interest. ACKNOWLEDGMENT This work was supported by the National Key Research & Development Program (Grant Nos. 2016YFA0201901 and 2016YFA0201902), the National Science Foundation of China (Grant Nos. 61621061, 61390504 and 61427901). References (1) Simrén, M.; Stotzer, P.O. Use and Abuse of Hydrogen Breath Tests. Gut 2006, 55, 297-303. (2) Hübert, T.; Boon-Brett, L.; Black, G.; Banach, U. Hydrogen Sensors-A Review. Sens. Actuators, B 2011, 157, 329-352. (3) Buttner, W.J.; Post, M.B.; Burgess, R.; Rivkin, C. An Overview of Hydrogen Safety Sensors and Requirements. Int. J. Hydrogen Energ. 2011, 36, 2462-2470. (4) Varghese, O.K.; Gong, D.; Paulose, M.; Ong, K.G.; Grimes, C.A. Hydrogen Sensing Using Titania Nanotubes. Sens. Actuators, B 2003, 93, 338-344. (5) Favier, F.; Walter, E.C.; Zach, M.P.; Benter, T.; Penner, R.M. Hydrogen Sensors and Switches From Electrodeposited Palladium Mesowire Arrays. Science 2001, 293, 2227-2231. (6) Hughes, R.C.; Schubert, W.K. Thin Films of Pd/Ni Alloys for Detection of High Hydrogen Concentrations. J. Appl. Phys. 1992, 71, 542-544. (7) Yang, F.; Kung, S.; Cheng, M.; Hemminger, J.C.; Penner, R.M. Smaller is Faster and More Sensitive: The Effect of Wire Size On the Detection of Hydrogen by Single Palladium Nanowires. ACS Nano 2010, 4, 5233-5244. 13 / 21


ACS Sensors 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(8) Kong, J.; Franklin, N.R.; Zhou, C.W.; Chapline, M.G.; Peng, S.; Cho, K.J.; Dai, H.J. Nanotube Molecular Wires as Chemical Sensors. Science 2000, 287, 622-625. (9) Kong, J.; Chapline, M.G.; Dai, H.J. Functionalized Carbon Nanotubes for Molecular Hydrogen Sensors. Adv. Mater. 2001, 13, 1384-1386. (10)

Mubeen, S.; Zhang, T.; Yoo, B.; Deshusses, M.A.; Myung, N.V. Palladium Nanoparticles Decorated Single-Walled Carbon Nanotube Hydrogen Sensor. J. Phys. Chem. C 2007, 111, 6321-6327.

(11)

Sun, Y.G.; Wang, H.H. High-Performance, Flexible Hydrogen Sensors that Use Carbon Nanotubes Decorated with Palladium Nanoparticles. Adv. Mater. 2007, 19, 2818-2823.

(12)

Khalap, V.R.; Sheps, T.; Kane, A.A.; Collins, P.G. Hydrogen Sensing and Sensitivity of Palladium-Decorated Single-Walled Carbon Nanotubes with Defects. Nano Lett. 2010, 10, 896-901.

(13)

Ganzhorn, M.; Vijayaraghavan, A.; Dehm, S.; Hennrich, F.; Green, A.A.; Fichtner, M.; Voigt, A.; Rapp, M.; von Löhneysen, H.; Hersam, M.C.; Kappes, M.M.; Krupke, R. Hydrogen Sensing with Diameter- And Chirality-Sorted Carbon Nanotubes. ACS Nano 2011, 5, 1670-1676.

(14)

Zhang, M.L.; Brooks, L.L.; Chartuprayoon, N.; Bosze, W.; Choa, Y.; Myung, N.V. Palladium/Single-Walled Carbon Nanotube Back-to-Back Schottky Contact-Based Hydrogen Sensors and their Sensing Mechanism. ACS Appl. Mater. Inter. 2014, 6, 319-326.

(15)

Li, X.W.; Le Thai, M.; Dutta, R.K.; Qiao, S.P.; Chandran, G.T.; Penner, R.M. Sub-6 Nm Palladium Nanoparticles for Faster, More Sensitive H2 Detection Using Carbon Nanotube Ropes. ACS Sensors 2017, 2, 282-289.

(16)

Sun, Y.G.; Wang, H.H.; Xia, M.G. Single-Walled Carbon Nanotubes Modified with Pd Nanoparticles: Unique Building Blocks for High-Performance, Flexible Hydrogen Sensors. J. Phys. Chem. C 2008, 112, 1250-1259.

(17)

Du, Y.G.; Xue, Q.Z.; Zhang, Z.Y.; Xia, F.J.; Liu, Z.L.; Xing, W. Enhanced 14 / 21


Page 14 of 21

Page 15 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Sensors

Hydrogen Gas Response of Pd Nanoparticles-Decorated Single Walled Carbon Nanotube Film/SiO2/Si Heterostructure. AIP Adv. 2015, 5, 027136. (18)

Lee, J.; Kang, W.; Najeeb, C.K.; Choi, B.; Choi, S.; Lee, H.J.; Lee, S.S.; Kim, J. A Hydrogen Gas Sensor Using Single-Walled Carbon Nanotube Langmuir–Blodgett Films Decorated with Palladium Nanoparticles. Sens. Actuators, B 2013, 188, 169-175.

(19)

Wongwiriyapan, W.; Okabayashi, Y.; Minami, S.; Itabashi, K.; Ueda, T.; Shimazaki, R.; Ito, T.; Oura, K.; Honda, S.; Tabata, H.; Katayama, M. Hydrogen Sensing

Properties

of

Protective-Layer-Coated

Single-Walled

Carbon

Nanotubes with Palladium Nanoparticle Decoration. Nanotechnology 2011, 22, 055501. (20)

Samanta, S.K.; Fritsch, M.; Scherf, U.; Gomulya, W.; Bisri, S.Z.; Loi, M.A. Conjugated Polymer-Assisted Dispersion of Single-Wall Carbon Nanotubes: The Power of Polymer Wrapping. Accounts Chem. Res. 2014, 47, 2446-2456.

(21)

Homenick, C.M.; Rousina-Webb, A.; Cheng, F.; Jakubinek, M.B.; Malenfant, P.R.L.; Simard, B. High-Yield, Single-Step Separation of Metallic and Semiconducting SWNTs Using Block Copolymers at Low Temperatures. J. Phys. Chem. C 2014, 118, 16156-16164.

(22)

Sridharan, R.; Mahendran, K.H.; Nagaraj, S.; Gnanasekaran, T.; Periaswami, G. An Electrochemical Hydrogen Meter for Measuring Hydrogen in Sodium Using a Ternary Electrolyte Mixture. J. Nucl. Mater. 2003, 312, 10-15.

(23)

Gu, J.T.; Han, J.; Liu, D.; Yu, X.Q.; Kang, L.X.; Qiu, S.; Jin, H..H; Li, H.B.; Li, Q.W.; Zhang, J. Solution-Processable High-Purity Semiconducting SWCNTs for Large-Area Fabrication of High-Performance Thin-Film Transistors. Small 2016, 12, 4993-4999.

(24)

Zhong, D.; Zhang, Z.Y.; Ding, L.; Han, J.; Xiao, M.M.; Si, J.; Xu, L.; Qiu, C.G.; Peng, L.M. Gigahertz Integrated Circuits Based On Carbon Nanotube Films. Nature Electronics 2018, 1, 40.

(25)

Ishikawa, F.N.; Curreli, M.; Olson, C.A.; Liao, H.; Sun, R.; Roberts, R.W.; 15 / 21


ACS Sensors 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Cote, R.J.; Thompson, M.E.; Zhou, C. Importance of Controlling Nanotube Density for Highly Sensitive and Reliable Biosensors Functional in Physiological Conditions. ACS Nano 2010, 4, 6914-6922. (26)

Kim, B.; Park, N.; Na, P.S.; So, H.; Kim, J.; Kim, H.; Kong, K.; Chang, H.; Ryu, B.; Choi, Y. The Effect of Metal Cluster Coatings On Carbon Nanotubes. Nanotechnology 2005, 17, 496.

(27)

Kauffman, D.R.; Star, A. Chemically Induced Potential Barriers at the Carbon Nanotube − Metal Nanoparticle Interface. Nano Lett. 2007, 7, 1863-1868.

(28)

Bhowmick, R.; Rajasekaran, S.; Friebel, D.; Beasley, C.; Jiao, L.Y.; Ogasawara, H.; Dai, H.J.; Clemens, B.; Nilsson, A. Hydrogen Spillover in Pt-Single-Walled Carbon Nanotube Composites: Formation of Stable C-H Bonds. J. Am. Chem. Soc. 2011, 133, 5580-5586.

(29)

Roland, U.; Braunschweig, T.; Roessner, F. On the Nature of Spilt-Over Hydrogen. J. Mol. Catal. A: Chem. 1997, 127, 61-84.

(30)

Lin, C.K.; Yang, Z.Z.; Xu, T.; Zhao, Y.F. An in Situ Electrical Study On Primary Hydrogen Spillover From Nanocatalysts to Amorphous Carbon Support. Appl. Phys. Lett. 2008, 93, 233110.

(31)

Koo, W.; Qiao, S.; Ogata, A.F.; Jha, G.; Jang, J.; Chen, V.T.; Kim, I.; Penner, R.M. Accelerating Palladium Nanowire H2 Sensors Using Engineered Nanofiltration. ACS Nano 2017, 11, 9276-9285.

16 / 21


Page 16 of 21

Page 17 of 21

Table 1 Comparison of key metrics for CNT based H2 sensors of this work with other representative published works to date. Material

Detection limit

Pd /CNT(single tube)

9

Response

40ppm

2

Pd /CNT(single tube)12

100ppm

Pd@CNT rope15

Response

(400ppm)

time

Scalability

5-10s

No

1000 (300ppm)

480s

No

10ppm

0.95 (1800ppm)

710s

Yes

Pd/SiO2/CNTnetwork19

1ppm

0.29 (100ppm)

10s

Yes

Pd@CNTnetwork10

100ppm

1.5 (500ppm)

18min

Yes

Pd@CNTnetwork11

100ppm

0.7 (500ppm)

15s

Yes

Pd@s-CNTnetwork

0.89ppm

~1000 (311ppm)

7s

Yes

(This work)

1.2 S33

M11

b

r-CNT s-CNT

S22

1.0

Raman intensity (a.u.)

a Absorption

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Sensors

0.8 0.6 0.4 0.2

G

r-CNT s-CNT

RBM D

0.0 600

c

800 1000 1200 Wavelength (nm)

1400

150

d

200 1200 1400 1600 1800 -1 Raman shift (cm )

Figure 1 Characteristics of CNT materials used for building sensors. Absorption spectrum (a) and Raman spectrum (b) of sorted CNT (green) and raw CNT (blue). SEM picture (c) and AFM image (d) of the as-deposited s-CNTs films. The height profile in down inset of (d) shows the thickness of the film is ~2 nm, indicating a monolayer-like film. 17 / 21


ACS Sensors

b

a

c

e

d

1E-5 1E-5

s-CNT

1E-6 1E-7

1E-7

311ppmH2

Ids(A)

Ids(A)

1E-8

1E-10

1E-9 1E-10

s-CNT s-CNT with Pd NPs r-CNT r-CNT with Pd NPs

1E-11 1E-12

1E-11 -60

1E-13 -40

-20

0

Vgs(V)

20

40

60

H2on

1E-7

N2

N2

1E-9

1E-8

H2off

Air

1E-8

1000ppm H2

0

200

r-CNT+Pd NP s-CNT+Pd NP only Pd NP 400

600

800

Time(s)

1E-5

r-CNT

H2off

1E-6

1E-6

Ids(A)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 21

∆I

1E-6

∆Vgs

H2on

1E-9

1E-7

1E-10

-60 -30 0 30 Vgs(V)

60

-60 -30 0 30 Vgs(V)

60

Figure 2 Structure and preliminary characteristics of CNT-film H2 sensors. (a) Optical image of one unit consisting 24 sensors. Inset: optical image of the as-fabricated sensors array. (b) Schematic diagram of the Pd-decorated CNT film H2 sensor. Lower inset is the SEM image of 1 nm Pd-decorated CNT channel. (c) Transfer curves of H2 sensors based on s-CNT(olive) and r-CNT (blue) without (line) and with (circle) Pd decoration. Vds=-0.1 V. (d) Time dependent response to 311 ppm H2 of s-CNT (olive) , r-CNT (blue) and Pd NP without CNTs (red) based sensors. (e) Transfer curves of the sensors based on s-CNT (left) and r-CNT (right) before (blue) and after (red) 311 ppm H2 exposure for 200 s.

18 / 21


Page 19 of 21

20

1200 16

b 20ppm

311ppm

1400

15ppm

1200

0.89ppm 7.5ppm 2ppm 4 8

900

0

600

×3

0

206ppm

Response

Response

12

×3

155ppm 104ppm 83ppm 62ppm 40ppm

600 1200 1800 2400 Time(s)

300

1000 800 600 400

-0.01

200

-0.02

0

0 0

30

60

90

120

0

100

200 300 Time(s)

150

Time(min)

c

1000ppm CH4 1000ppm CO2

0.00 (R-R0)/R0

1500 24

(R-R0)/R0

a

0

d

100

200

300

100

1000

Concentration(ppm)

Pd@ s-CNT network(this work)

1000

1E-6

Pd@ r-CNT network(this work) Pd contact single s-CNT12 Pd@single s-CNT9

τresponse=7s

100

τrecovery=89s

1E-8

1E-9

0

1000

2000

3000

Pd@SiO2/r-CNT network

4000

10

Pd@r-CNT rope15 Pd@r-CNT network10

1 0.1

Device I Device II Fitting curve

1E-10

Pd contact alligned r-CNT14 Pd@r-CNT network11

Response

1E-7

Ids(A)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Sensors

0.01 5000

1

10

Concentration(ppm)

Time(s)

Figure 3 Characteristics of CNT film based H2 sensors. (a) Real time response to different H2 concentration at room temperature. Inset: the magnified image under low H2 concentration. (b) Relationship between response and H2 concentration. Inset shows dinky response to 1000 ppm CH4 and CO2. Vds=-1V. (c) Real time response of two H2 sensors with repeated three on/off sensing cycles under H2 concentration of 311 ppm for 200 s. Response time and decay time can be defined by fitting portions of the curves (Blue line) to exponentials of the form e-t/τ . Vds= -0.1V. (d) Gas response comparison of several previous reported best CNT based hydrogen sensors.

19 / 21


ACS Sensors

a

b

c

d 1000

Room Temprature ~100°°C

100 Response

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 21

10

1

0.1 0.1

1

10

100

Concentration(ppm)

Figure 4 Mechanism exploration of CNT film based H2 sensors. (a) Equivalent circuit scheme of a CNT network H2 sensor. (b) Schematic of H2-Pd/s-CNT interaction in the channel. (c) Band diagram of different stage during H2 sensing. As shown here, doping in the channel will also influence the contact resistance by changing the barrier thickness near the contact. (d) Response of gas sensors to different H2 concentrations at different operation temperatures.

20 / 21


Page 21 of 21

Table of contents 24 20

1200 16

1E-5

Sorted CNT film H2off

1E-6

H2on

1E-7

1E-8

Raw-CNT film H2off

Response

12 8

900

4 0

1E-5

×3

1E-6

206ppm

×3

155ppm 104ppm 83ppm 62ppm 40ppm

600 1200 1800 2400 Time(s)

∆I ∆Vgs

311ppm

15ppm

0.89ppm 7.5ppm 2ppm

0

600

20ppm

(R-R 0)/R0

1500

Ids(A)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Sensors

300

H2on

1E-9

0 1E-7

1E-10

-60 -30 0 30 Vgs(V)

60

-60 -30 0 30 Vgs(V)

60

0

30

60

90

120

150

Time(min)

21 / 21


Batch Fabrication of Ultrasensitive Carbon Nanotube Hydrogen

Recommend Documents