Home
Search
Collections
Journals
About
Contact us
My IOPscience
A stretchable and highly sensitive chemical sensor using multilayered network of polyurethane nanofibres with self-assembled reduced graphene oxide
This content has been downloaded from IOPscience. Please scroll down to see the full text. 2017 2D Mater. 4 025062 (http://iopscience.iop.org/2053-1583/4/2/025062) View the table of contents for this issue, or go to the journal homepage for more Download details: IP Address: 128.138.73.68 This content was downloaded on 04/04/2017 at 03:51 Please note that terms and conditions apply.
2D Mater. 4 (2017) 025062
https://doi.org/10.1088/2053-1583/aa6783
PAPER
RECEIVED
26 January 2017 RE VISED
13 March 2017
A stretchable and highly sensitive chemical sensor using multilayered network of polyurethane nanofibres with self-assembled reduced graphene oxide
ACCEP TED FOR PUBLICATION
17 March 2017 PUBLISHED
3 April 2017
Le Thai Duy1, Tran Quang Trung1, Adeela Hanif1, Saqib Siddiqui1, Eun Roh2, Wonil Lee3 and Nae-Eung Lee1,2,3 1
School of Advanced Materials Science and Engineering, Sungkyunkwan University, Suwon, Kyunggi-do 16419, Republic of Korea SKKU Advanced Institute of Nanotechnology (SAINT), Sungkyunkwan University, Suwon, Kyunggi-do 16419, Republic of Korea 3 Samsung Advanced Institute for Health Sciences and Technology (SAIHST), Sungkyunkwan University, Suwon, Kyunggi-do 16419, Republic of Korea 2
E-mail:
[email protected] Keywords: multilayer, reduced graphene oxide, polyurethane nanofibre, gas sensor, stretchable electronics Supplementary material for this article is available online
Abstracts Stretchable chemical sensors for detection of gases/vapours are of great interest for wearable applications in healthcare and environmental safety. Here, we demonstrate a simple but effective approach based on layering electro-spun elastomeric polyurethane (PU) nanofibres to produce stretchable chemically responsive nanohybrids of PU nanofibres with self-assembled reducedgraphene oxide (R-GO). The study using a chemiresistive structure reveals the positive effects of layering the elastomeric fibres of R-GO/PU nanohybrids on both stretching and sensing capabilities. The R-GO/PU devices layered four and five times own excellent sensitivity towards NO2 (50 ppb‒5 ppm at room temperature) as well as mechanical sustainability under static and cyclic stretching tests up to 50% of strain. These stretchable chemiresistors based on multilayering of hybrid fibre network of an elastomer and 2D sensing nanomaterials provides a good premise for wearable chemical sensing applications.
1. Introduction Monitoring the environment is important for protecting people from pollutants and toxic gases continually increasing nowadays. Chemical sensors, particularly gas sensors, which can detect and monitor dangerous matters in gas phase at very low concentrations beyond human senses are of increasing necessity for environment safety. Recently, rapid growth of modern technology, especially the appearance of nanomaterials with sensing devices for wearable electronics and emerging internet of things (IoT) applications, has opened new opportunities to improve quality of life as well as enable everyone to monitor adverse changes in the environment and be aware of risks anytime and anywhere [1, 2]. Although rigid chemical sensors have been well developed and widely used, there are many wearable applications requiring devices possessing good flexibility or stretchability for their integration onto nonplanar surfaces (e.g. clothes, papers, plants, and animals) for real-time monitoring [3–5]. © 2017 IOP Publishing Ltd
Along these lines, there have been many studies on development of flexible sensing nanomaterials as well as cost-effective fabrication of high-performance flexible sensors that can sustain mechanical deformations like bending [5–17], and strategies to tune their sensing performance [8, 11, 12, 18–21]. Although the probability of integrating flexible chemical sensors in practical wearable applications is very high, many cases demand stretchable sensing platforms which are capable of accommodating and sustaining a larger deformation range than bending. For example, good sustainability to a larger deformation range and stretching is required for electronic skin (e-skin) which can be used for replacing injured human skin (especially of nose and tongue) or used for inspection robots in harsh conditions. Another case is integration of chemical sensing devices into or on textiles and clothing (easily elongated when dressing, taking off, washing, and even while using in daily activities or sports like swimsuits and socks) to detect body odours or explosive/toxic matters in the environment. Besides, other on-body applications like smart gloves (for doctors, farmers, and industrial workers) and
2D Mater. 4 (2017) 025062
L T Duy et al
bandages or patches (for monitoring humidity, odours of wounds) as well as food safety applications like smart wraps/bags (for spoiling detection) also require stretchable chemical sensing devices. However, development of stretchable chemical sensors has been very hindered due to difficulty in obtaining the mechanical sustianability of chemical sensing materials and maintaining sensing capability under large mechanical deformation. To fabricate the expected stretchable electronic devices, it is critical that the employed materials should be deformable, prolongable, and friendly to users and the environ ment. Many intrinsically stretchable materials for wearable sensing applications have been developed recently [1, 2, 22–28], but most of them are better used as electrodes or in physical sensing applications where sensitivity to chemical analytes is not required. Among stretchable sensing materials, reduced graphene oxide (R-GO), which is non-toxic, high conductive, and sensitive to dangerous matters at room temperature, is a promising candidate widely studied for low-cost wearable chemical sensors. Since stacked or assembled networks of R-GO nanosheets have shown limited stretchability, there have been some strategies developed for boosting stretchability of R-GO up to 30% or higher such as pre-stretching [26, 29–31], usage of stretch-inducing substrates [32], and stretchable patterns [33–35]. However, each approach has its own disadvantages. These include: (i) need for encapsulation (which is not suitable for gas sensing materials) to prevent delamination, (ii) need for increased substrate thickness that leads to low attachability on curvilinear surfaces, (iii) complex and expensive equipment to make stretchable patterns at sub-microns scales, and (iv) difficulty in fabricating multifunctional devices (having multiple mat erial depositions, alignments, and annealing steps). Thus, many issues need to be overcome to enable practical usage of these strategies including ease of mass production and the fabrication of low-cost, lightweight wearable devices. Apart from creating a microarchitecture for the sensing area or sensing material, compositing or hybridizing R-GO nanosheets with elastomeric materials is another potential choice. There have been many reports on highly stretchable physical sensing devices based on R-GO/elastomer composites or hybrids such as strain sensors [30, 36–39], temperature sensors [40, 41], etc. These reports show that compositing and hybridizing are highly promising ways to obtain stretchable R-GO-based materials at low cost to expand the functions of wearable sensing platforms such as e-skin and sportswear (>50% strain). However, there are few studies on developing stretchable composites or hybrid materials for chemical sensing [42, 43]. There are many challenges that remain before these composites or hybrid materials can be used in practical applications including: the limited permeability of gas molecules for capture and release of target gases inside the composite, and hindered reaction area due to low reactive components in a composite, resulting in low sensitivity to analytes at sub part-per-million (ppm) concentration [44]. Additionally, mechanisms and effects of methods for obtaining stretchability on 2
sensing performance have not been fully investigated. Such studies are necessary to expand the functions of wearable sensing platforms [45]. Overall, developing stretchable chemical sensing materials still needs more research to enable integration of stretchable chemical sensors into practical wearable sensing applications. In this work, we demonstrate a simple but controllable approach to achieve a highly stretchable transparent chemiresistive materials formed via a series of electrospinning polyurethane (PU) nanofibres in orthogonal directions and subsequent self-assembly of R-GO nanosheets on the nanofibre network (i.e. multiple layering or multilayering). The R-GO/PU nanohybrid fibres were formed directly on hydrophobic polydimethylsiloxane (PDMS) substrate without the need for any additional wetting processes. Throughout the use of a chemiresistive platform, demonstrative gas sensing measurements were carried out to determine the sensing performance of stretchable nanohybrid R-GO/PU fibre sensors to NO2 (50 ppb–5 ppm) at room temperature. Besides, the effects of the fibre density and multilayering on sensing performance as well as stretching sustainability were also studied in detail. With four and five times of layering, the sensors based on R-GO/PU nanohybrid fibres sustained more than 5000 cycles at 30% elongation plus 5000 cycles at 50% elongation. Finally, the stretching and sensing measurements in the synthetic environment and ambient conditions showed that multilayering nanohybrid R-GO/PU fibres is promising to obtain stretchable chemical sensors with high performance for smart wearable sensing applications, especially healthcare and environment safety.
2. Experiments 2.1. Fabrication of materials and devices Details related to the fabrication process of all materials and devices (including back-gated transistors and chemiresistors) are described in supporting information section 1 (stacks.iop.org/TDM/4/025062/ mmedia). 2.2. Characterization and sensing measurements FESEM and Raman data were captured by using JEOL JSM-6500F and Alpha300R (WITec) equipment, respectively. I–V characteristics were measured with HP 4145B (Agilent) and Keithley 2400 equipment. Electrical gas sensing data were recorded with a source meter (Keithley 2400) at a constant applied voltage of 5 V under a fixed total flow (1000 sccm) of dry air and gases. The sensing system is demonstrated in supporting information section 1.
3. Results and discussion 3.1. Fabrication and characterization Our target in obtaining stretchable chemiresistors with high mechanical sustainability and sensing performance was to investigate the effects of increased
2D Mater. 4 (2017) 025062
L T Duy et al
(a)
(b)
GaIn electrodes
(c) Transmittance (%)
PDMS substrate
100 80 60 1. PDMS 2. R-GO/PU (5) 3. (5\3) 4. (5\3\3) 5. (8\3\3\1) 6. (8\3\3\3\1)
40 20 0 200
400
600
800
Wavelength (nm)
Figure 1. (a) Schematic structure of a device based on multilayered R-GO/PU nanofibre network. (b) and (c) Optical image and UV–visible spectra show optical transparency of a PDMS sample (numbered as 1) and others numbered from 2 to 6 corresponding to the R-GO/PU samples (5), (5\3), (5\3\3), (8\3\3\1), and (8\3\3\3\1), respectively.
electrical pathways and higher sensing area that were achieved by fabricating multilayer-networked fibres of R-GO/PU nanohybrids. To perform electrical tests, stretchable gas sensing devices in a simple resistive configuration as demonstrated in figure 1(a) were prepared. Here, using an adhesive elastomer like PU is very useful. First, electrospun PU nanofibres can stick directly to the hydrophobic PDMS substrate. Then, a monolayer assembly of graphene oxide (GO) nanosheets on the PU nanofibre network also could take place quickly (within 5−10 min) without any pre-treatment. That the process does not require any surface treatment is potential to avoid damaging or contaminating materials existing on the surfaces/ substrates like our multilayered fibre network. After one step of PU electrospinning and subsequent one step of GO adsorption, one GO-covered-PU fibre network (i.e. ‘layer’) was formed. To increase electric pathways in multiple directions, the samples at the odd turns of electrospinning (1st, 3rd, and 5th) were placed in parallel with PU fibres (y-direction) while they at the even turns of electrospinning (2nd and 4th) were rotated 90° clockwise (x-direction) as shown in figure 1((a)-bottom) and supporting figure S1.1. Besides, the electrospinning time and the number of layers were varied to study the effects of the number of layers as denoted in table 1. The next step is reduction of GO to form electrical conductive R-GO/PU channel. To minimize damage to PU fibres (especially melting of nanofibres by temperatures above 100 °C, supporting figure S1.2), chemical reduction of GO by hydrazine vapour (35% in DI) at 50 °C was carried out with a long reduction time (18 h). This allowed the hydrazine vapour to effectively penetrate inside the nanofibre network. After reduction of GO, the samples were still highly transparent as displayed in figure 1(b). 3
Here, our approach makes these materials suitable for transparent wearable applications with aesthetic or private purposes. Although increasing the number of layers obviously brought about reduced visiblelight transmittance (from 76 to 65% at a wavelegnth of 550 nm as shown in figure 1(c)), the R-GO/PU sample (8\3\3\3\1) is still highly transparent with more than 60% transmittance in the visible light region due to the transparency of PU and nanometrescale thickness of R-GO nanosheets (about 3−5 nm of single flakes and 5−15 nm of stacking flakes checked by atomic force microscopy (AFM) shown in supporting figure S1.3). In order to study the effects of multilayering of the nanohybrid fibres, we used the stretchable interdigitated electrodes made from a metallic liquid (Gallium based alloy). This allowed us to explore maximum performance of the hybrid materials without considering limitations associated with contact resistance or low electrode conductivity under stretching tests [46–48]. To coat highly conductive and stretchable GaIn alloy liquid, a buffer layer of 5 nm of Ti and 20 nm of Au was deposited by e-beam evaporation through a stencil mask with 400 µm gap (Chartpak drafting and design applique film, patterned by a laser cutter). The same stencil mask was continually used for patterning of the GaIn electrode, which was deposited by using a soft brush tool. The complete device after removing the stencil mask is illustrated in figure 1(a). Field-emission scanning electron microscopy (FESEM) and Raman measurements were performed to check the self-assembly of GO and R-GO on PU nanofibres and the potential damage of PU nanofibres after the reduction process. The FE-SEM images in figures 2(a)–(c) show the PU and R-GO/PU fibres stacked in the multilayered network for one of the samples (8\3\3\3\1). Comparing these FE-SEM images, we
2D Mater. 4 (2017) 025062
L T Duy et al
Table 1. Notation for the samples/devices based on R-GO and PU fibres fabricated under different conditions of electrospinning time and turns. Electrospinning time (min) 1st
2nd
3rd
4th
5th
Sample name
3
—
—
—
—
R-GO/PU (3)
5
—
—
—
—
8
—
—
—
—
(8)
14
—
—
—
—
(14)
3
3
—
—
—
(3\3)
5
3
—
—
—
(5\3)
5
3
3
—
—
(5\3\3)
5
5
3
—
—
(5\5\3)
5
3
3
1
—
(5\3\3\1)
5
3
3
3
—
(5\3\3\3)
8
3
3
1
—
(8\3\3\1)
8
3
3
3
—
8
3
3
3
1
(8\3\3\3\1)
8
3
3
3
3
(8\3\3\3\3)
8
5
3
3
1
—
—
—
—
—
(5)
(8\3\3\3)
(8\5\3\3\1) R-GO only
observed similar diameters of both PU and R-GO/PU fibres (about 0.8‒0.9 µm), but there was a clear change in surface roughness between them (under the same reduction and annealing conditions). This suggests that the coverage of GO (R-GO) nanosheets on PU fibres was successful and that PU fibres survived the GO reduction process. The Raman spectra of PU, GO/PU, and R-GO/PU fibres in figure 2(d), showing the Raman peaks of PU have no change, verify that the PU fibres were not damaged by reduction process of GO. Furthermore, the intensity change in D and G peaks (1344 and 1596 cm−1, respectively) of the GO and R-GO samples confirmed the successful reduction process [11, 48]. For further understanding electrical properties of the nanohybrid materials and the effects of PU fibres on R-GO properties, back-gated transistors based on R-GO only and R-GO/PU (8) were fabricated (as described in supporting information section 1). The data in supporting figure S1.4 revealed that the R-GO and R-GO/PU channels are the same hole-rich type because their neutral charge points (i.e. Dirac point, VDirac) are in positive gate voltage (about 4‒6 V). Besides, there is no clear difference in their VDirac values, indicating that there was no significant effect of PU fibres (electrical insulator) on the electrical properties of R-GO. 3.2. Effects of fibre networking conditions on stretchability and sensing characteristics To study the effects of fibre networking conditions, current intensities proportional to the number of electrical pathways were monitored for comparisons. The voltage across interdigitated electrodes would be fixed at 5 V. Stretching tests along y direction (illustrated in figure 3(a) where a device was elongated 30−50%) were conducted for all devices to compare gas sensing performance and to understand effects of multilayering 4
or increasing current pathways because the PU fibres were firstly and majorly electrospun in y direction (demonstrated in figure 1(a)). To differentiate the causes for the current changes, the normalized current (I/Io, where Io is the initial current intensity and I is the dynamic current intensity measured under stretching tests) was used for strain, while the current response (ΔI/I (%) = 100 × (Igas − Iair)/Iair where Iair and Igas are the currents stabilized in dry air and during gas response tests, respectively) was used for gas exposures. Here, strain is ε = 100% × (L1 − L0)/L0; L1 and L0 are the device lengths in the y-direction after and before stretching, respectively. Before investigating the effects of fibre networking conditions, determining sensing behaviour of the R-GO material is also necessary for understanding changes in current. Therefore, a representative sensing measurement of a chemiresistive device based on only R-GO was carried out (in dry air, at room temper ature, with a bias voltage of 5 V, ε = 0%, and the system setup is mentioned in supporting figure S1.5). As shown in figure 3(b), the device conductance increases upon exposures of NO2 (oxidizing gas), while decreasing towards NH3 (reducing gas) exposures. This result indicates that hole is the main charge carrier in the R-GO materials because while NH3 molecule is an electron donor leading to the increased recombination of electrons and holes (i.e. reduced charge carriers) in the sensing channel, NO2 molecule is electron acceptor leading to the decreased charge recombination as well as increased conductivity of the device. In other words, our R-GO materials are p-type semiconductors, which are in agreement with the data obtained by the back-gated devices (supporting figure S1.4). Furthermore, the device responded to 2.5 ppm NO2 very fast and seemed to reach a stable responsivity (ΔI/I ~ 40%) under 3 min NO2 exposures (4 pulses) while its responses to 25 ppm NH3 gas were slower (ΔI/I = −15 to −20% for 3 min exposure) and about two NH 3 pulses were required to bring the current down to the baseline, indicating a higher and faster responsivity of the R-GO towards NO2 gas. This is probably because the low thermal reduction process (at 90 °C to avoid damage of PU) could not remove amine groups completely (from hydrazine) and thus leave more electronrich sites in the R-GO network, resulting in the more attraction of electron-withdrawing molecules like NO2 [50–52]. Consequently, NO2 was mainly used for sensing measurements and comparisons. Besides, the gas exposure time and purging time would be fixed at 2 min (fast response region) and 15 min, respectively. To estimate the device stretchability, current–volt age (I–V) measurements were carried out under one static stretching test (i.e. test A with static elongation strain εA from 0 to 50%) and two cyclic stretching tests. Test B is the first cyclic stretching test with 5000 cycles at elongation 30%. Test C is the second cylic stretching test with additional 5000 cycles at elongation 50%. And gas sensing measurements at static elongation strains of εB
2D Mater. 4 (2017) 025062
L T Duy et al
(a) R-GO/PU fibres
(b) PU fibres
(d)
D
Intensity (a.u.)
(c) R-GO/PU fibres
(1)
(2)
(1) PU fibres (2) GO/PU (3) R-GO/PU
G
(3)
2D
0
1000
2000
S3
3000
Raman shift (cm-1) Figure 2. (a) FE-SEM image of multilayered R-GO/PU fibre network in sample (8\3\3\3\1). (b) and (c) FE-SEM images of PU fibres before and after R-GO self-assembly, respectively. (d) Raman spectra of PU, GO/PU, and R-GO/PU samples. Here, the letters of D, G, 2D, S3 are corresponding to Raman bands of graphene materials.
and εC after test B and test C, respectively, were also carried out. Representative I–V data of the R-GO/PU fibre device (8\3\3\3\1), having a highly stable sensing signal and stretching performance, is shown in figure 3(c). After extracting mean current values, the normalized current data (I/Io) were plotted as a function of strain (ε) which is displayed in figure 3(d). Here, figures 3(c) and (d) show that with increasing the elongation from 0 to 50% in test A, its current dropped from 1.1 to 0.035 µA (I/Io = 0.032, at ±5 V). After this test, the device did not fully recover to the initial value (I/Io = 1.0 → 0.42). To understand what happened inside the stretched channel as well as in the fibre network, we used an optical microscope to monitor morphological changes. After observing and comparing the changes either in the open channel region or at fibre/fibre junctions (in figures S2.1 and S2.2—supporting information section 2, respectively), we found that the fibres were elongated in y-direction and compressed in x-direction. They clearly followed the deformation of the substrate due to the high adhesion. Moreover, many cracks were observed on the surface of the PDMS elastomer substrate (supporting figure S2.2(b)–(d)), indicating that elastic PU fibres could be damaged under large elongations. Although these optical images do not reveal whether there was any damage of fibres such as cracking, breaking or sliding at the junctions, a significant change in the angle at fibre intersection was clearly 5
observed. These observations reveal that the fibre junctions and R-GO self-assembled layers could have experienced damages at the nanometre scale. To confirm the damage of the R-GO networked layer which was considered as reconnection problem of R-GO nanosheets, a test based on thermal relaxation was carried out. Here, there were many overlapped and/or stacked regions of R-GO nanosheets when they were self-assembled on PU fibres, resulting in the high roughness of the R-GO/ PU fibres as shown in figure 2(c). Under stretching, the R-GO nanosheets were strained, and the overlapped R-GO nanosheets could be separated. The discontinuities in the R-GO networked layer may look like cracks on the PDMS surface and can lead to reduction in the the current under elongation. When the strain was released, the separated R-GO nanosheets could be reconnected, restoring the current pathways. Obviously, not all R-GO nanosheets could be overlapped again because wrinkles of R-GO nanosheets could be formed after the strain release, leading to the reduced current pathways compared to the initial condition. Thus, we tried to release the wrinkles by thermal relaxation at 90 °C for 2−6 hr because PU fibres do not melt at this temperature. After 2 hr annealing, an improved conductance recovery in the device was observed as shown in figure 3(c) (I/Io = 0.42 → 0.65). Here, the thermal relaxation may not effectively recover the changes at the fibre/fibre junctions (especially fibre
2D Mater. 4 (2017) 025062
L T Duy et al
(a)
Current, I (µA)
50
50 40 30
45
20 40
GAS ON
GAS OFF
10 0
35 30 25 200
NO2 exposures
-10
NH3 exposures
-20
300
400
500
3000
6000
(c)
0.4
(3) After annealed
40% 50%
0.0 1.0
-0.1
0.0 -0.5
-0.3 -0.4
(2) -6
(3) -4
0.9 0.4 0.3
(1)
Re
lea
0.2
-2
(e)
sed
(2) 10
20
30
40
Response (%)
Response (%)
(1) (2)
50
1000
(1) No stretch after test B (2) At 30% stretch (3) At 50% stretch 2000
3000
6
(2) (3) 100
(1)
50
(1) No stretch (2) At 30% stretch (3) At 50% stretch 0
(g)
150
0
4
1000
2000
3000
4000
5000
Time (s)
After 1st cyclic stretching test (test B)
(3)
5
2
150
50
200
0
0
During test A
Strain (%)
100
-5
200
0
0
(f)
0
0.1 0.0
(3)
-1.0
(1)
Voltage (V)
Before cyclic test (1) After 1st cyclic test (2) After 2nd cyclic test
Response (%)
Normalized Current, I/Io
Initial 1.1 (I ~ 1.1 µA) 1.0 o
(1)
0.5
Time (s) 1.2
(d)
20% 30%
0.1
-0.2
9000 12000
strain = 10%
(1) Before stretch
0.3 (2) Release after stretch 0.2
Current (µA)
Fast response region (~ 2 min)
Response, ∆I/Io (%)
55
(b)
4000
5000
200
After 2nd cyclic stretching test (test C) 150
(2)
(1) 100
50
(1) No stretch after test C (2) At 30% stretch (3) At 50% stretch
(3)
0 0
1000
2000
3000
4000
5000
Time (s)
Time (s)
Figure 3. (a) Optical images of the R-GO/PU device (8\3\3\3\1) before (left), during 30% elongation (middle) and 50% elongation (right) loaded in a custom-built cyclic stretching tester. (b) Sensing responses of a rigid R-GO device towards exposures of 2.5 ppm NO2 and 25 ppm NH3 diluted in dry air. (c) I–V characteristics of the R-GO/PU device (8\3\3\3\1) before, during, and after statically stretched for the first time (test A). (d) Strain-dependent current changes (I/I0) of the R-GO/PU device under stretching tests. (e)–(g) Sensing responses of the R-GO/PU device towards 2.5 ppm NO2 gas at static strains (0, 30, and 50%) in test A, after test B and then test C, respectively. Here, test B includes 5000 stretching cycles at 30% elongation and test C includes 5000 stretching cycles at 50% elongation.
sliding) due to the good adhesion of PU. Consequently, it is confirmed that imperfect connections between R-GO nanosheets self-assembled on PU fibres were associated with the reduction of conductivity as well as the poor recovery. After annealing for a longer time (up to 6 h), the device conductance did not return to near the initial value, indicating permanent damage possibly at fibre junctions or intersections. As shown in figure 3(d), after performing test B (5000 cycles at 30% strain) and test C (5000 cycles 6
at 50% strain) in sequence for the R-GO/PU device (8\3\3\3\1), the conductance decreased to I/Io ≈ 0.154 (~0.17 µA) and then to I/Io ≈ 0.105 (~0.106 µA), respectively. A decrease in conductance at various strains (10 to 50%) was also observed after these cyclic tests, but the observed decreases were smaller than that of test A. Besides, the current recoveries after sequential cyclic tests of B and C (I/Io ≈ 0.154 → 0.136 and 0.105 → 0.088, respectively) are also better than that observed after test A (I/Io = 1 → 0.42). Although the
2D Mater. 4 (2017) 025062
L T Duy et al
conductance was significantly reduced as the result of further damage in the fibre network, device (8\3\3\3\1) still had a good enough conductance for sensing measurements towards 2.5 ppm NO2 gas at room temper ature as shown in figures 3(e)–(g). Here, immediately after each I–V test at a certain strain ε = 0, 30, and 50%, we performed one gas sensing measurement. In other words, figures 3(e)–(g) display the gas sensing data measured at static strains in the tests A, B, and C, respectively. Comparing the device responses towards 2 min exposures of NO2 at ε = 0, 30, and 50% before and after the cyclic tests, some small changes in current response can be observed due to the changes inside the R-GO network (e.g. disconnection in y-direction but reconnection in x-direction). In general, however, the sensing behaviour and average response magnitudes were maintained well (ΔI/I ≈ 100% / 2 min exposure), indicating the good sensing capability of the device. Although the representative data of the (8\3\3\3\1) device are presented in figures 3(e)–(g), we also compared the sensing and stretching capabilities for the R-GO/PU samples shown in table 1 to study the effects of fibre density and multilayering in the R-GO/PU fibre network. The measured data of other devices are shown in supporting information section 3, and the information necessary for comparison is summarized in supporting table S3.1. In the gas sensing data, we chose to compare the sensing value for the third NO2 exposure because some devices had been exposed to gas only three times. Besides, the gas responsivity of most devices after three pulses is close to the saturation/stable response (i.e. the highest response the device could reach). First, data associated with the samples with one R-GO/PU fibre layer are shown in supporting figure S3.1. The R-GO/PU device (3) owned a too low conductance (Io ≈ 8.5 nA) to obtain a clear gas sensing signal (i.e. transducing signal). Device (5) showed a better conductance (Io ≈ 0.3 µA) and responded well to NO2 gas before stretching (ΔI/I ≈ 35%). At εA = 30%, the conductance of device (5) is too low (Io ≈ 300 → 6 nA), resulting in a poor gas sensing signal. However, after released (εA = 0%), the sensing signal became clear again (ΔI/I ≈ 50%), but it was higher than that before it was stretched. In our opinion, the response difference was probably caused by changes in R-GO/PU fibre network, which was damaged and reconstructed under the strain as aforementioned. For device (8) with a better conductance (Io ≈ 0.6 µA), an increase in noise level that was similar to but smaller than that of device (5) was observed between 0% and 30% strain. The NO2 sensing signal was also higher (ΔI/I ≈ 60%). Although device (8) survived after test B, its signal degraded due to low conductance (Io ≈ 600 → 16 nA). For device (14) with the highest conductance (Io ≈ 1.5 µA), its gas responsivity (ΔI/I ≈ 120%) was significantly higher than those of devices (5) and (8) (ΔI/I ≈ 35 and 60%, respectively). After test B, its sensing signal was interfered by noise at εB = 30% (Io ≈ 1500 → 11 nA) while that of the device (8) was degraded at εB = 0%. In 7
general, the increase of electrospinning time as well as fibre density provided not only more current pathways (mainly in the y-direction), but also more sensing area, resulting in the improved sustainability to stretch and NO2 responsivity. To understand effects of increasing fibres in the x-direction, we compared the samples with one and two R-GO/PU layers. Data for these samples with two R-GO/PU layers are presented in supporting figure S3.2. By comparing device (3\3) to device (3) (Io ≈ 30 and 8.5 nA, respectively) and device (5\3) to device (5) (Io ≈ 0.5 and 0.3 µA, respectively), higher current intensities were clearly observed for devices (3/3) and (5/3), indicating enhancement of current pathways by adding fibres in the x direction. More interestingly, although the total electrospinning time as well as fibre density between devices (8) and (5\3) were rather similar, device (5\3) had a smaller number of current pathways in the y-direction, but it showed a better stretching and sensing capability (passed test B, ΔI/I ≈ 120 at εA = 30%) than that of device (8) (did not passed test B, ΔI/I ≈ 60 at εA = 30%) because device (5/3) had more pathways in the x direction. Nevertheless, between devices (3\3) and (5\3), device (5\3) had a higher fibre density and better stretch sustainability, but its gas responses were not higher than those of (3\3). Some similar situations were observed among the devices having three, four, and five R-GO/PU layers as shown in supporting figures S3.3–S3.5, respectively and mentioned below. Despite that unclear relationship between the increased current pathways with gas sensing enhancement, the improvement in stretching and sensing capability of the samples with two R-GO/ PU fibre layers illustrates the positive effects of increasing electrical pathways in the x direction in the hybrid R-GO/PU network. Next, to understand effects of multilayering, more comparisons were carried out among the samples having similar total electrospinning times, including R-GO/PU devices (8)-(5\3)-(5\3\1), (14)-(5\5\3)(5\3\3\3)-(8\3\3\1), and (8\3\3\3)-(8\3\3\3\1). The three-layer samples survived after test B, while the four-layer ones passed test C and the five-layered ones sustained all tests with elongation strain εA,B,C = 0 to 50% as shown in supporting table S3.1. These results confirmed the enhancement in stretching sustainability realized by multilayering the fibre network. About gas sensing data, although the sensing signals and noise interference in devices (5\3\1) and (5\3) were similar, device (5\3\1) had a higher response than that of device (5\3). Both of these devices were more responsive to gas than device (8). This indicates that stacking R-GO/ PU fibres may have reduced contact area between fibres and PDMS substrate, leading to the increased sensing area. Among devices (14), (5\5\3), (5\3\3\3) and (8\3\3\1) (Io ≈ 1.5, 1.5, 1.4, and 1.1 µA, respectively), device (14) had the lowest stretchability while device (8\3\3\1) with the lowest conductance showed the best sustainability to stretch. Regarding gas sensing
2D Mater. 4 (2017) 025062
L T Duy et al
c apability, device (5\5\3) showed a clearer transducing signal at εB = 30% but a low magnitude (ΔI/I ≈ 70%), which was even lower than that of devices (14) and (5\3\1) (ΔI/I ≈ 120% and 145% at εB = 30%, espectively). Device (8\3\3\1) had a better sustainability at εC = 30% that was similar to device (5\3\3\3), but its sensing responses (ΔI/I ≈ 160%) were lower than that of device (5\3\3\3) (ΔI/I ≈ 210% at εC = 30%). Regarding devices (8\3\3\3) and (8\3\3\3\1), device (8\3\3\3) also survived at εC = 50% but its gas sensing signal was a bit interfered by increased noise level at this strain. Although the five-layered samples, including (8\3\3\3\1), had a higher fibre density and stretching sustainability, their conductivity (Io ⩽ 1.9 µA) and gas responsivity (ΔI/I ≈ 110 to 150%) were not better than those of device (8\3\3\3) (Io ≈ 2.0 µA and ΔI/I ≈ 140 to 160%). In additional, the gas responses of devices (8\3\3\3) were not better than those of devices (5\3\3\1) and (5\3\3\3) (ΔI/I ≈ 160 to 210%). These comparisons again show a non-proportional relationship between increase in the electrical pathways by multilayering and enhancement in gas sensing. These results may come from differences not only in the number of R-GO/PU fibres across the sensing channel and fibre/fibre junctions (i.e. overlapped sensing area) but also in contact area of fibre/PDMS in the channel exposed to gas. Overall, the obtained data and comparisons confirmed the enhancement in sustainability to stretch and maintenance of sensing signal under high strains via multilayering the R-GO/PU nanohybrid network. 3.3. Gas sensing performance of stretchable R-GO/PU fibre sensors towards NO2 For practical sensing applications, the gas sensors or sensing materials must be able to detect toxic gases at concentrations lower than limit values which can cause adverse effects on health. For chemical substances, there are three types of threshold limit values (TLV), such as time-weighted average (TWA, allow 8 h exposure a day and 40 h/week), short-term exposure limit (STEL, allow 15 min/exposure and a maximum of 4 exposures/day with at least 60 min intervals), and a ceiling value (should not exceed at any time) that have been researched by professional associations like the American Conference of Governmental Industrial Hygienists (ACGIH), the Occupational Safety and Health Administration (OSHA), and the National Institute for Occupational Safety and Health (NIOSH). Currently, the recommended TWA for NO2 gas is 0.2 ppm by ACGIH, the STEL is 1 ppm by NIOSH, and the ceiling value is 5 ppm by OSHA [53–56]. Therefore, for practical usage, the sensors should be capable of detecting NO2 gas in a large range, at lease from 0.2 to 5 ppm. That is the reason why the responsivity of devices towards various NO2 concentrations (2 min/exposure) at room temperature are presented to illustrate that the sensing performance of the hybrid R-GO/PU material is sufficient for use in practical applications. Here, 8
the sensing data of the R-GO/PU device (8\3\3\3\1) towards various NO2 concentrations (from 0.05 to 5 ppm) are shown in figure 4. For further comparison, additional sensing measurements were also performed with devices including R-GO only, R-GO/PU (14), and R-GO/PU (5\3\3\1) as displayed in figure S4.1 in supporting information section 4. Regarding the self-assembled R-GO device on glass substrate (supporting figure S4.1(a)), its current level was very high (Io ≈ 30 µA at 5 V) but its gas responsivity was much lower (ΔI/I ≈ from 7.5% at 0.05 ppm to 38% at 2.5 ppm NO2) than that of other devices. In our opinion, although the R-GO device had a number of abundant current pathways (due to its large scale continuous networked layer), there could be many regions as well as electrical highways not exposed to gas, such as overlapped areas of R-GO nanosheets shown in the AFM image of supporting figure S1.3(c)). This can be the cause of the small current changes (Igas − Iair), and thus the lower responses (ΔI/Iair). Regarding the R-GO/PU devices (14) and (5\3\3\1), their conductances after stretching tests were dropped a lot, but they still responded well to very low NO2 concentrations (50 ppb–2.5 ppm) as shown in supporting figures S4.1(b) and (c). As shown, the response of device (14) after test B was from 4.0% to 125% at 0.05–2.5 ppm, and that of device (5\3\3\1) after test C is from 8.5% to 205% at at 0.05–2.5 ppm. For device (8\3\3\3\1), the sensing responses checked at the beginning of test A and after cyclic stretching tests are as shown in figures 4(a) and (b), respectively. With increasing strain, further current decreases and changes in the sensing signal magnitude were observed in figures 4(b)–(d). The initial gas responsiveness of device (8\3\3\3\1) was from 12.9% at 0.1 ppm, 145% at 2.5 ppm to 163% at 5 ppm. After test C, its responsiveness at εC = 0% was from 3.5% at 0.05 ppm, 137% at 2.5 ppm to 167% at 5 ppm. The responses of device (8\3\3\3\1) at εC = 30% was from 6.8% at 0.05 ppm, 14.3% at 0.1 ppm, 154% at 2.5 ppm to 183% at 5 ppm. At εC = 50%, its responses were from 5.6% at 1 ppm, 144% at 2.5 ppm to 176% at 5 ppm. Although unexpected changes in responses occurred under strain tests due to the possible reasons as aforementioned, the high overall performance of the R-GO/PU device (8\3\3\3\1) towards ultralow NO2 concentrations was revealed after more than total 10 000 stretching cycles. Here, the theoretical noise and limit of detection (LOD) were also calculated in a manner similar to that presented in our previous work [57, 58], and the results are shown in supporting table S4.3. The R-GO device had the smallest noise level (ΔI/I ≈ 1.11 × 10−3% due to the very high conductance) and thus its LOD towards NO2, about 0.055 ± 0.008 ppb with a response resolution of 0.61%/10 ppb, was the lowest (best). R-GO/ PU devices had higher noise and LOD values before or after stretching tests than those of the R-GO device. The noise and LOD values of device (14) were about 4.05 × 10−2% and 1.45 ± 0.07 ppb, repectively. Values
2D Mater. 4 (2017) 025062
0.8
3.5
90
0.4
3.0 2.5
120
60
0.2
1.5
0.1
4000
7000
10000
50
0.4
0.1 0.2 0 4000
7000
Current (µA)
0.25
5 ppm 2.5
150
1.6 0.20
0.15
0.8
0.1
100
0.4
0.05
(d)
200
50
0.2
0.10
0 1000
4000
7000
10000
0.18
13000
At 50% stretch after test C
13000
Time (s)
180
2.5
0.14
150
1.6
0.12
120
0.8
0.10
90
0.4
0.08 0.06
210
5 ppm
0.16
Current (µA)
At 30% stretch after test C
Response (%)
0.30
10000
Time (s)
Time (s)
(c)
100
0.8 0.05
1000
13000
150
1.6
0.2
0 1000
0.3
5 ppm 2.5
0.4
30
2.0
200
No stretch after test C 0.5
150
1.6
4.0
0.6
Response (%)
2.5
4.5
(b)
180
5 ppm
0.1
60
90% recovery ~ 2.5 hr 30
0.2
1000
Response (%)
Current (µA)
5.0
Before stretch in A test
Current (µA)
5.5
Response (%)
(a)
L T Duy et al
0 4000
7000
10000
13000
16000
Time (s)
Figure 4. (a) Sensing responses of the R-GO/PU device (8\3\3\3\1) at 0% elongation in test A towards various NO2 concentrations (0.1–5 ppm) at room temperature. (b) Sensing responses of the device at 0% elongation after two cyclic stretching tests B and C towards NO2 gas (50 ppb–5 ppm). (c) and (d) NO2 concentration dependent responses of the device at static 30% and 50% elongations after test C, respectively.
of device (5\3\3\1) were 4.30 × 10−2% and 0.90 ± 0.06 ppb. Values of device (8\3\3\3\1) before stretching were 1.92 × 10−2% and 0.60 ± 0.09 ppb, while those after test C were 2.19 × 10−2% and 0.91 ± 0.03 ppb. Especially here, the noise level of device (8\3\3\3\1) was the smallest (best) in comparison to devices (5\3\3\1) and (14) after stretching tests. While the gas sensitivity of device (8\3\3\3\1) was lower than that of the device (5\3\3\1), their LOD (=3 * noise/sensitivity) values are similar as the compensation result of noise level and sensitivity, indicating a potential of multilayering which is a possibility of the limit detection enhancement due to increased current pathways and thus reduced signal noise. Besides, the R-GO/PU devices were fabricated with short electrospinning times as well as low density of fibres for studying the effects of fibre networking conditions. Therefore, further multilayering of the nanohybrid fibre network and an increase in fibre density can more improve not only stretching sustainability but also noise level as well as LOD. Figures 4(c) and (d) and supporting table S4.3 show that the noise level of device (8\3\3\3\1) increased gradually but LOD values did not with increasing elongation (εC = 0–50%). In our opinion, the LOD of the device (8\3\3\3\1) at εC = 30% (0.73 ± 0.06 ppb) was better than that at εC = 0% (0.91 ± 0.03 ppb) and 50% (1.07 ± 0.07 ppb) because the increased noise level and variated sensitivity were compensated for by changes in the fibre network. Despite changes in the theoretical data at different strains, these data helped to confirm 9
the high sensitivity of our material. Compared to sensing peformance of other flexible NO2 sensors based on graphene materials as shown in table 2, the obtained results of our R-GO/PU devices indicate the possibility of applying these devices to wide range detection and monitoring of NO2 gas. Unfortunately, the high sensitivity due to more gas attraction and associated reactions could come along with slow gas desorption, and thus, it took 2.5–3 h to recover by 90%–95%, as shown in figure 4(d) and supporting figure S4.1(c). Anyway, based on numerous reported studies of chemical sensors, this issue of R-GO materials can be solved by, for example, functionalization, hybridization with other nanomaterials, and so on. This also is a good premise for further studies on improving stretchable sensing materials. Finally, to understand the effects of environements on the sensing capabilities and possibility of differentiating gas response from strain response of device (8\3\3\3\1), one sensing measurement (demonstrative for smart gloves) in ambient conditions (setup and measurement descriptions are in supporting information section 5) was performed and recorded into a video clip (supplementary video SV) by using a portable camera. Gas responses of the device towards two pulses of 1 ppm NO2 gas (40 s/pulse; the STEL is 15 min/pulse) are clearly and quickly seen, and it may be possible to differentiate these from strain signals caused by finger movements (combination of stretching and bending) as shown in supporting figure S5.1. However, the
2D Mater. 4 (2017) 025062
L T Duy et al
Table 2. Comparison of sensing performance of flexible NO2 sensors based on graphene materials. Sensor response
Ref.
Sensing materials
Response (%/ ppm/min)
Recovery (min)
Range (ppm)
Bending/Stretching sustainability (cycles)a
[6]
Vertical CNTs/R-GO
14/5/60
>60
0.5–10
[17]
R-GO based yarn
6/0.25/30
>400
0.1–5
1000 (r = 1 mm)
[59]
graphene/MoS2
5/5/5
30
1.2–5
5000 (r = 1.9 mm)
[60]
Ag-S/R-GO
74.6/50/0.2
0.33
0.5–50
1000 (r = 1 cm)
[32]
R-GO on mogul-pattern
65/25/1
12
2.5–25
2000 (ε = 30%)
This work
Multilayered R-GO/PU fibres
90/1/2
150
0.05–5
5000 (ε = 50%)
n.a. (r = 15 mm)
Bending radius (flexible devices) = r; elongation strain (stretchable devices) = ε.
a
obtained data from this measurement reveal that the strain responses were still rather high and there was a coeffect of temperature, humidity, and pressure on NO2 responses. The co-effect was confirmed by additional measurements under different humidity, temperature, and mass flow conditions in supporting figure S5.2. Through observations from the supplementary video and additional data, our device performance still needs more improvement to minimize effects of strain and surrounding environments as well as increase the selectivity before applied to practical stretchable sensing applications. Anyway, with huge exposed R-GO surface as well as sensing space, our hybrid approach, compared to embedding R-GO inside other elastomers, still is the good potential for further studies that involve functionalization and hybridization with other nanomaterials to enhance stretchability, sensitivity and selectivity.
4. Conclusion In summary, our stretchable transparent sensing materials based on nanohybrid of R-GO nanosheets and PU fibres were successfully developed by a multilayering strategy to open more current pathways which can enable more sustainability to stretch and maintenance of sensing signal under high strains. Stretchable chemical sensors were obtained by using a chemiresistive structure with stretchable metallic liquid electrodes. The effects of fibre density and multilayering of the R-GO/PU nanohybrid fibre network were studied by adjusting electrospinning time and directions, and the results indicated that the increase in fibre networking could enhance stretch sustainability, noise as well as detection limit of the hybrid material. The target stretchability (50%) for wearable applications could be obtained by the devices with four and five R-GO/PU layers. Furthermore, some sensing measurements towards NO2 at a large concentration ranges (0.05‒5 ppm) and a measurement demonstrating the utility of the materials for wearable applications at ambient conditions were carried. The obtained data showed the potential of using our device as well as hybrid sensing material for developing practical wearable sensing applications. Finally, our study reveals a promising approach to achieve better 10
stretchable chemical sensing materials and shows that there are still rooms for improving and applying these materials for other stretchable applications via further hybridization or functionalization.
Acknowledgments This research was supported by the Basic Science Research Program (Grant No. 2016R1A2A1A05005423) through the National Research Foundation (NRF) funded by the Ministry of Science, ICT& Future Planning.
Author information The authors declare no competing financial interests.
References [1] Bao Z and Chen X 2016 Flexible and stretchable devices Adv. Mater. 28 4177–9 [2] Hussain A M and Hussain M M 2016 CMOS-technologyenabled flexible and stretchable electronics for internet of everything applications Adv. Mater. 28 4219–49 [3] Kim J-H, Chun J, Kim J W, Choi W J and Baik J M 2015 Self-powered, room-temperature electronic nose based on triboelectrification and heterogeneous catalytic reaction Adv. Funct. Mater. 25 7049–55 [4] Lee K, Park J, Lee M-S, Kim J, Hyun B G, Kang D J, Na K, Lee C Y, Bien F and Park J-U 2014 In situ synthesis of carbon nanotube–graphite electronic devices and their integrations onto surfaces of live plants and insects Nano Lett. 14 2647–54 [5] He Q, Sudibya H G, Yin Z, Wu S, Li H, Boey F, Huang W, Chen P and Zhang H 2010 Centimeter-long and large-scale micropatterns of reduced graphene oxide films: fabrication and sensing applications ACS Nano 4 3201–8 [6] Jeong H Y, Lee D-S, Choi H K, Lee D H, Kim J-E, Lee J Y, Lee W J, Kim S O and Choi S-Y 2010 Flexible roomtemperature NO2 gas sensors based on carbon nanotubes/ reduced graphene hybrid films Appl. Phys. Lett. 96 213105 [7] Ryu G-S, Park K H, Park W-T, Kim Y-H and Noh Y-Y 2015 High-performance diketopyrrolopyrrole-based organic fieldeffect transistors for flexible gas sensors Org. Electron. 23 76–81 [8] Pandey S, Goswami G K and Nanda K K 2013 Nanocomposite based flexible ultrasensitive resistive gas sensor for chemical reactions studies Sci. Rep. 3 2082 [9] Yao K, Caruntu D, Wozny S, Huang R, Ikuhara Y H, Cao B, O’Connor C J and Zhou W 2012 Towards one key to one lock: catalyst modified indium oxide nanoparticle thin film sensor array for selective gas detection J. Mater. Chem. 22 7308–13 [10] Kinkeldei T, Zysset C, Münzenrieder N and Tröster G 2013 The influence of bending on the performance of flexible carbon black/polymer composite gas sensors J. Polym. Sci. B 51 329–36
2D Mater. 4 (2017) 025062
L T Duy et al
[11] Duy L T, Trung T Q, Dang V Q, Hwang B-U, Siddiqui S, Son I-Y, Yoon S K, Chung D J and Lee N-E 2016 Flexible transparent reduced graphene oxide sensor coupled with organic dye molecules for rapid dual-mode ammonia gas detection Adv. Funct. Mater. 26 4329–38 [12] Rashid T-R, Phan D-T and Chung G-S 2014 Effect of Ga-modified layer on flexible hydrogen sensor using ZnO nanorods decorated by Pd catalysts Sensors Actuators B 193 869–76 [13] Ammu S, Dua V, Agnihotra S R, Surwade S P, Phulgirkar A, Patel S and Manohar S K 2012 Flexible, all-organic chemiresistor for detecting chemically aggressive vapors J. Am. Chem. Soc. 134 4553–6 [14] Asad M, Sheikhi M H, Pourfath M and Moradi M 2015 High sensitive and selective flexible H2S gas sensors based on Cu nanoparticle decorated SWCNTs Sensors Actuators B 210 1–8 [15] Bai S, Sun C, Wan P, Wang C, Luo R, Li Y, Liu J and Sun X 2015 Transparent conducting films of hierarchically nanostructured polyaniline networks on flexible substrates for highperformance gas sensors Small 11 306–10 [16] Choi S-J, Kim S-J and Kim I-D 2016 Ultrafast optical reduction of graphene oxide sheets on colorless polyimide film for wearable chemical sensors NPG Asia Mater. 8 e315 [17] Yun Y J, Hong W G, Choi N-J, Kim B H, Jun Y and Lee H-K 2015 Ultrasensitive and highly selective graphene-based single yarn for use in wearable gas sensor Sci. Rep. 5 10904 [18] Niu S, Hu Y, Wen X, Zhou Y, Zhang F, Lin L, Wang S and Wang Z L 2013 Enhanced performance of flexible ZnO nanowire based room-temperature oxygen sensors by piezotronic effect Adv. Mater. 25 3701–6 [19] He Q, Zeng Z, Yin Z, Li H, Wu S, Huang X and Zhang H 2012 Fabrication of flexible MoS2 thin-film transistor arrays for practical gas-sensing applications Small 8 2994–9 [20] Shukla S, Agrawal R, Cho H J, Seal S, Ludwig L and Parish C 2005 Effect of ultraviolet radiation exposure on roomtemperature hydrogen sensitivity of nanocrystalline doped tin oxide sensor incorporated into microelectromechanical systems device J. Appl. Phys. 97 054307 [21] Lee J S, Jun J, Shin D H and Jang J 2014 Urchin-like polypyrrole nanoparticles for highly sensitive and selective chemiresistive sensor application Nanoscale 6 4188–94 [22] Jang H, Park Y J, Chen X, Das T, Kim M-S and Ahn J-H 2016 Graphene-based flexible and stretchable electronics Adv. Mater. 28 4184–202 [23] Yan C, Cho J H and Ahn J-H 2012 Graphene-based flexible and stretchable thin film transistors Nanoscale 4 4870–82 [24] Choi S, Lee H, Ghaffari R, Hyeon T and Kim D-H 2016 Recent advances in flexible and stretchable bio-electronic devices integrated with nanomaterials Adv. Mater. 28 4203–18 [25] Chortos A and Bao Z 2014 Skin-inspired electronic devices Mater. Today 17 321–31 [26] Ahn J-H and Je J H 2012 Stretchable electronics: materials, architectures and integrations J. Phys. D: Appl. Phys. 45 103001 [27] Lee H, Kim M, Kim I and Lee H 2016 Flexible and stretchable optoelectronic devices using silver nanowires and graphene Adv. Mater. 28 4541–8 [28] Trung T Q and Lee N-E 2016 Flexible and stretchable physical sensor integrated platforms for wearable human-activity monitoringand personal healthcare Adv. Mater. 28 4338–72 [29] Chae S H et al 2013 Transferred wrinkled Al2O3 for highly stretchable and transparent graphene–carbon nanotube transistors Nat. Mater. 12 403–9 [30] Liu Q, Chen J, Li Y and Shi G 2016 High-performance strain sensors with fish-scale-like graphene-sensing layers for fullrange detection of human motions ACS Nano 10 7901–6 [31] Zang J, Ryu S, Pugno N, Wang Q, Tu Q, Buehler M J and Zhao X 2013 Multifunctionality and control of the crumpling and unfolding of large-area graphene Nat. Mater. 12 321–5 [32] Lee H-B, Bae C-W, Duy L T, Sohn I-Y, Kim D-I, Song Y-J, Kim Y-J and Lee N-E 2016 Mogul-patterned elastomeric substrate for stretchable electronics Adv. Mater. 28 3069–77
11
[33] Vatani M, Lu Y, Lee K-S, Kim H-C and Choi J-W 2013 Direct-write stretchable sensors using single-walled carbon nanotube/polymer matrix J. Electron. Packag. 135 011009 [34] Yun J, Lim Y, Jang G N, Kim D, Lee S-J, Park H, Hong S Y, Lee G, Zi G and Ha J S 2016 Stretchable patterned graphene gas sensor driven by integrated micro-supercapacitor array Nano Energy 19 401–14 [35] Kim K, Kim J, Hyun B G, Ji S, Kim S-Y, Kim S, An B W and Park J-U 2015 Stretchable and transparent electrodes based on in-plane structures Nanoscale 7 14577–94 [36] Cheng Y, Wang R, Sun J and Gao L 2015 A stretchable and highly sensitive graphene-based fiber for sensing tensile strain, bending, and torsion Adv. Mater. 27 7365–71 [37] Hu D, Wang Q, Yu J, Hao W, Lu H, Zhang G, Wang X and Qiu L 2016 Highly stretchable strain sensors using an electrospun polyurethane nanofiber/graphene composite J. Nanosci. Nanotechnol. 16 5839–42 [38] Tang Y, Zhao Z, Hu H, Liu Y, Wang X, Zhou S and Qiu J 2015 Highly stretchable and ultrasensitive strain sensor based on reduced graphene oxide microtubes–elastomer composite ACS Appl. Mater. Interfaces 7 27432–9 [39] Lin Y, Liu S, Chen S, Wei Y, Dong X and Liu L 2016 A highly stretchable and sensitive strain sensor based on grapheneelastomer composites with a novel double-interconnected network J. Mater. Chem. C 4 6345–52 [40] Trung T Q, Ramasundaram S, Hwang B-U and Lee N-E 2016 An all-elastomeric transparent and stretchable temperature sensor for body-attachable wearable electronics Adv. Mater. 28 502–9 [41] Hua C, Shang Y, Li X, Hu X, Wang Y, Wang X, Zhang Y, Li X, Duan H and Cao A 2016 Helical graphene oxide fibers as a stretchable sensor and an electrocapillary sucker Nanoscale 8 10659–68 [42] Park J et al 2016 Wearable, wireless gas sensors using highly stretchable and transparent structures of nanowires and graphene Nanoscale 8 10591–7 [43] Kim J, Lee M-S, Jeon S, Kim M, Kim S, Kim K, Bien F, Hong S Y and Park J-U 2015 Highly transparent and stretchable fieldeffect transistor sensors using graphene–nanowire hybrid nanostructures Adv. Mater. 27 3292–7 [44] Kim H, Miura Y and Macosko C W 2010 Graphene/ polyurethane nanocomposites for improved gas barrier and electrical conductivity Chem. Mater. 22 3441–50 [45] Ding F, Ji H, Chen Y, Herklotz A, Dörr K, Mei Y, Rastelli A and Schmidt O G 2010 Stretchable graphene: a close look at fundamental parameters through biaxial straining Nano Lett. 10 3453–8 [46] Hirsch A, Michaud H O, Gerratt A P, de Mulatier S and Lacour S P 2016 Intrinsically stretchable biphasic (solid– liquid) thin metal films Adv. Mater. 28 4507–12 [47] Joshipura I D, Ayers H R, Majidi C and Dickey M D 2015 Methods to pattern liquid metals J. Mater. Chem. C 3 3834–41 [48] Boley J W, White E L, Chiu G T-C and Kramer R K 2014 Direct writing of gallium–indium alloy for stretchable electronics Adv. Funct. Mater. 24 3501–7 [49] Moon I K, Lee J, Ruoff R S and Lee H 2010 Reduced graphene oxide by chemical graphitization Nat. Commun. 1 73 [50] Eigler S, Grimm S, Enzelberger-Heim M, Müller P and Hirsch A 2013 Graphene oxide: efficiency of reducing agents Chem. Commun. 49 7391–3 [51] Park S et al 2012 Chemical structures of hydrazine-treated graphene oxide and generation of aromatic nitrogen doping Nat. Commun. 3 638 [52] Chua C K and Pumera M 2013 Chemical reduction of graphene oxide: a synthetic chemistry viewpoint Chem. Soc. Rev. 43 291–312 [53] ACGIH 2015 Adopted threshold limit values 2015 TLVs and BEIs Based on the Document of the Threshold Limit Values for Chemical Substances and Physical Agents and Biological Exposure Indices (Cincinnati, OH: American Conference of Governmental Industrial Hygienists (ACGIH)) pp 1–64
2D Mater. 4 (2017) 025062
L T Duy et al
[54] US Dep. Labor. OSHA—chemical sampling information | nitrogen dioxide. Available from: https://www.osha.gov/dts/ chemicalsampling/data/CH_257400.html [Accessed 14th Nov 2016] [55] US Dep. Labor. Limits for air contaminants—occupational safety and health administration. Available from: https:// www.osha.gov/pls/oshaweb/owadisp.show_document?p_ table=STANDARDS&p_id=9992 [Accessed 14th Nov 2016] [56] Cent. Dis. Control Prev. Nitrogen dioxide—the registry of toxic effects of chemical substances | CDC/NIOSH. Available from: http://www.cdc.gov/niosh-rtecs/qw958940.html [Accessed 14th Nov 2016] [57] Dua V, Surwade S P, Ammu S, Agnihotra S R, Jain S, Roberts K E, Park S, Ruoff R S and Manohar S K 2010 All-
12
organic vapor sensor using inkjet-printed reduced graphene oxide Angew. Chem., Int. Ed. Engl. 49 2154–7 [58] Duy L T, Kim D-J, Trung T Q, Dang V Q, Kim B-Y, Moon H K and Lee N-E 2015 High performance three-dimensional chemical sensor platform using reduced graphene oxide formed on high aspect-ratio micro-pillars Adv. Funct. Mater. 25 883–90 [59] Cho B et al 2015 Chemical sensing of 2D graphene/ MoS2 heterostructure device ACS Appl. Mater. Interfaces 7 16775–80 [60] Huang L, Wang Z, Zhang J, Pu J, Lin Y, Xu S, Shen L, Chen Q and Shi W 2014 Fully printed, rapid-response sensors based on chemically modified graphene for detecting NO2 at room temperature ACS Appl. Mater. Interfaces 6 7426–33
