A stretchable and highly sensitive chemical sensor ...

96 downloads 346 Views 5MB Size Report
3 Samsung Advanced Institute for Health Sciences & Technology (SAIHST), Sungkyunkwan. University, Suwon, Kyunggi-do ..... Figure S3.1. Sensing data of the ...
SUPPORTING INFORMATION

A stretchable and highly sensitive chemical sensor using multilayered network of polyurethane nanofibres with selfassembled reduced graphene oxide

Le Thai Duy,1 Tran Quang Trung,1 Adeela Hanif,1 Saqib Siddiqui,1 Eun Roh,2 Wonil Lee,3 and Nae-Eung Lee 1,2,3,* 1

School of Advanced Materials Science and Engineering, Sungkynkwan University, Suwon,

Kyunggi-do 16419, Korea. 2

SKKU Advanced Institute of Nanotechnology (SAINT), Sungkyunkwan University, Suwon,

Kyunggi-do 16419, Korea. 3

Samsung Advanced Institute for Health Sciences & Technology (SAIHST), Sungkyunkwan

University, Suwon, Kyunggi-do 16419, Korea. * Email: [email protected]

SI.1

Supporting Information 1 1. Fabrication process for R-GO film and R-GO/PU nanohybrid fibre devices. a. Preparation of materials. A polydimethylsiloxane (PDMS, Sylgard 184 from Down Corning) solution was prepared by mixing part A (base) and part B (curing agent) at a ratio of 20:1. After degassing, the solution was cast onto a flat container, released for 30 min, and then annealed on a balanced hotplate at 80 oC for 3 hr, resulting in a formation of 1-mm-thick PDMS substrate. For electrospinning polyurethane (PU) fibres, PU particles (SG 85A from Teco Flex) were dissolved in dimethylformamide (DMF, Sigma Aldrich) solvent (15% in weight) at 90oC with stirring for 2 hr. To speed up the solvent evaporation during electrospinning, acetone (Sigma Aldrich) was added into the solution, resulting in a diluted PU solution with 12% in weight. Chemically exfoliated graphene oxide (GO) nanosheets was synthesized from graphite by the modified Hummer's method.[1] After that, they were dispersed in deionized (DI) water with a concentration of 1 mg.mL-1 by sonication for 2-hr and then centrifuging at 4000 rpm for 30 min.

b. Fabrication of R-GO devices. To form a chemiresistive R-GO device on a glass substrate, an electrostatic promotion layer of poly(diallyldimethylammonium chloride (PDDA 20% in water, Sigma Aldrich) was drop-cast on the substrate before the adsorption or self-assembly of GO flakes. After 15 min of GO adsorption, the sample was immerged and shaken in DI water. Then the sample was dried by N2 gas, resulting in the remains of GO flakes binding with PDDA, so-called a self-assembled monolayer. Subsequently, the sample was placed into a closed container with hydrazine (35% in water, Sigma Aldrich) on a 50oC hotplate to convert GO into RGO. After 18-hr reduction, the sample was rinsed with DI water, dried with N2, and then annealed at 90oC for 5 hr to remove the remaining hydrazine. Finally, the electrodes of Ti and Au were deposited via e-beam and

SI.2

thermal evaporations through a stencil mask with a 400 µm the channel length or finger gap (Chartpak applique film patterned by a laser cutter). To fabricate the back-gated R-GO device on a glass substrate, a back-gate electrode made of Cr\Ni (10\40 nm) was first deposited on the substrate via e-beam evaporation. A 100-nm dielectric layer of Al2O3 was formed via atomic layer deposition (ALD) at 200oC. After that, the same process of GO adsorption and reduction as described above was carried out. Finally, the source-drain electrodes of Ti\Au (10\30 nm) were deposited through a shadow mask with a smaller open channel (80 µm in length). The schematic device structure is shown in Figure S1.4(a). c. Fabrication of R-GO/PU nanohybrid fibre devices To fabricate multilayer-networked R-GO/PU fibre chemiresistive devices on PDMS substrate, PU fibres were electrospun from PU solution (12%) directly on the substrate placed under the nozzle about 15 cm) at a DC voltage of 15 kV as shown in Figure S1.1. Subsequently, the samples were released in air to evaporate the acetone/DMF solvent for a few minutes before GO adsorption. The GO dispersion was dropped on the fibre layer for 5 min, and the sample was then rinsed with DI water and dried with N2, resulting in the formation of one layer of PU fibres self-assembled with GO nanosheets. After that, electrospinning process was carried out again, but the sample was rotated 90oC. Following solvent evaporation for a few minutes, GO adsorption step was also repeated without any pretreatment, leading to the formation of the second layer of GO/PU fibres. These electrospinning and GO adsorption steps would be repeated to obtain more networked layers of GO/PU fibres. At the last layer, the time for the GO adsorption step was 10 min for GO flakes to ensure good coverage on the fibre network The reduction process with hydrazine vapour was performed at 50oC for 18 hr. After cleaning with DI water to remove the remaining hydrazine and annealed at 90oC for 4 hr, the samples were introduced into the chamber of the e-beam and thermal evaporation system for Ti and Au deposition (10\20nm) through a stencil mask with the same interdigitated electrode pattern of the R-GO resistive device. Finally, metallic liquid alloy (eutectic GaIn) was brushed on the Au layer through the used stencil mask, resulting in the formation of the highly stretchable conductive electrodes.

SI.3

To fabricate back-gated devices with R-GO/PU fibres (device (8)) on glass substrates, the depositions of Ni gate electrode and Al2O3 dielectric layer were performed in the same way as the R-GO transistor device. After the ALD process, the electrospinning (8 min) and GO adsorption steps (10 min) were carried out just one time and followed by the same hydrazine reduction. Finally, the source-drain electrodes were deposited by e-beam and thermal evaporation system via a shadow mask (80-µm channel length). The device structure is shown in Figure S1.4(b).

2. Supplementary data for characterization of stretchable R-GO/PU chemiresistor device

Figure S1.1. Device fabrication. (a) Schematic process of producing multilayered R-GO/PU hybrid fibre chemiresistor. (b) Direction of electrospun nanofibres in odd (in y-axis) and even number of layers (in x-axis). (c) Optical images of a sample R-GO/PU (8\3\3\3\1) before/after coated with GaIn liquid electrodes (left, scale bar = 1 cm) and a corner of the device channel with an electrode gap of 400 µm (right).

SI.4

Figure S1.2. A test to estimate melting temperature of R-GO/PU nanohybrid fibres. (a-b) Optical images of a networked elastomeric R-GO/PU fibres (after fabrication with a maximum temperature of 90 oC for 4 hr) at 10× and 100× magnification (focused area is indicated by a rounded-rectangle), respectively. (c,d) Images of networked R-GO/PU fibres that underwent post-annealing steps at 100oC for 1 hr and then 130oC for 30 min, respectively. Here, there was no clear melted or joined fibres that were observed after annealing at 100oC, but they were melted at 130oC, indicating the melting temperature of R-GO/PU fibres that could be above 100oC.

SI.5

(a)

(b)

(c)

(d)

Figure S1.3. Morphology of materials. (a) Atomic force microscopic (AFM) image of R-GO nanosheets spincoated on a SiO2 wafer. (b) Profiling of R-GO nanosheets (profile 1‒5) to determine thickness and length of RGO nanosheets obtained by Gwyddion, a free data analysis software. Here, based on observations of profiles 1, 4, and 5, the size and thickness of single R-GO nanosheets varied in the range of 0.1‒1 µm and 3‒5 nm, respectively. The increased thickness (7‒13 nm) at the stacked area of R-GO nanosheets was also seen in profiles 2 and 3. (c-d) AFM image and thickness information of a self-assembled R-GO layer on PDDA-treated SiO2 wafer, respectively. Here, the self-assembled R-GO network layer has many overlapped/stacked areas, and the roughness at these areas can be higher than 15 nm (about 3‒5 stacking folds).

SI.6

3. Characterization of back gated devices To measure current-voltage (I-V) characteristics of the back-gated devices, a semiconductor analyzer system HP 4145B (Aligent) was used with a fixed source-drain voltage (V DS = 2 V) and a sweeping gate voltage (VG = −10 to +20 V). Through analysis of the ambipolar characteristics of R-GO based devices, useful information relating to the main charge carrier as well as semiconducting behaviour were revealed. As shown in Figure S1.4, the source-drain current level (IDS) of the R-GO devices with a large continuous self-assembled R-GO areas were much higher than that of the R-GO/PU fibre device (8). Besides, the field-effect could be reduced by using insulating PU fibres located between the R-GO nanosheets and gate dielectric layer, resulting in a lower current level in R-GO/PU fibre devices. This indicates that the R-GO conductive pathways in the R-GO/PU fibre devices mainly formed on the PU fibre network. Furthermore, both of devices based on R-GO only and R-GO/PU fibres had positive neutral charge points (VDirac = 4.0 to 6.0 V), where the concentrations of hole (p-branch) and electron (n-branch) were equally induced by the gating effect. This indicates the p-type semiconducting behaviour of both RGO and R-GO/PU materials. Because the measurements were performed in ambient conditions and in a sequence, there may be certain effects from the surrounding environment, especially humidity and other gases/vapours, which can lead to small changes in their current level and position of VDirac. Overall, the data indicate the insignificant effect of the insulating PU material on R-GO electrical properties.

SI.7

(a)

(b)

21

VDirac

ch an br p-

Source-Drain Current (A)

22

c ran n-b

20

h

19

Source-Drain Current (A)

7.8

R-GO back-gated devices

23

R-GO/PU back-gated devices

7.6 7.4

VDirac

7.2 7.0 6.8 6.6 6.4 6.2

18 -10

-5

0

5

10

15

20

Gate Voltage (V)

-10

-5

0

5

10

15

20

Gate Voltage (V)

Figure S1.4. I-V characterization of back-gated devices. (a) Device structure and transfer curves of three R-GO devices fabricated in the same batch. (b) Device structure and transfer curves of four R-GO/PU fibre devices (8) fabricated in the same batch. The transfer curves were obtained at a sweeping gate voltage (VG = −10 to +20 V) and a fixed drain voltage (VDS = 2 V).

SI.8

4. System setup for gas sensing measurements

Figure S1.5. Gas sensing system. Here, samples were placed inside a glass tube (i.e., the sensing chamber). Air and gas flows (1000 sccm in total) were adjusted via mass flow controllers (MFC). The temperature of measurements was controlled/maintained by a furnace. A humidifier was used to adjust humidity level of the air flow. Keithley 2400 is a power supply and source meter. The fume hood was opened to allow for ventilation to obtain measurements relevant for practical applications. A 3D-printed zig, which can statically stretch sensor devices inside the measurement chamber, was used for gas sensing measurements of the devices under stretching.

SI.9

Supporting Information 2 Characterization of changes in the multilayer-networked R-GO/PU nanohybrid fibres in the device (8\3\3\3\1) under stretching

Figure S2.1. Changes in the multilayer-networked R-GO/PU fibre device (8\3\3\3\1). (a) Optical image of a corner in the device channel (with 400 µm gap) before stretching. (b-d) Optical images of the focused corner (scale bar = 400 µm) at 20, 30 and 50% stretch, respectively. Here, the device was elongated in ydirection.

SI.10

(a)

Before stretch

50 (b)

50 50

(c)

50 (d)

50

100 At 20% stretch

y 100 100 At 30% stretch

100 At 50% stretch

100

SI.11

(e)

Release

50

100

Figure S2.2. Changes in fibre junctions (a) Optical images of an area at magnification of 50× (left) and 100× (right) in the non-stretched sensing channel, respectively. (b-d) Optical images of the focused area stretched at 20, 30 and 50% in the y-direction, respectively. (e) Optical images of the focused area returned after stretching test. The scale bars in the left and right images are 20 and 10 μm, respectively.

SI.12

Supporting Information 3 Data of other R-GO/PU fibre devices (1 to 5 layers) obtained with a fixed voltage of 5 V.

Test A

0.8 0.6

Rele

as e

d

0.4 0.2

1.0

100

R-GO/PU (5)

Initial (Io ~ 0.3 A)

R-GO/PU (5)

Test A

0.8 0.6

Re

0.4

lea sed

0.2

(1) No stretch (2) At 30% stretch (3) Release

80 Test A

Response (%)

1.0

1.2

R-GO/PU (3)

Initial (Io ~ 8.5 nA)

Normalized Current, I/Io

Normalized Current, I/Io

1.2

60

(3) 40

20

(1) (2)

0

0.0

0.0 0

5

10

15

20

25

30

0

10

Strain (%)

Rel ea

sed

0.2

0

10

20

30

80

40

40

(2) 20

50

2000

3000

4000

0

5000

lea

sed

0.2

20

30

Strain (%)

2000

150

40

50

4000

5000

Time (s) R-GO/PU (14) After test B

Test A

150

(3) 100

3000

200

(2) (1)

50

(1) No stretch (2) At 30% stretch (3) Release

0

After 1st cyclic test 10

No stretch 1000

1000

2000

3000

Time (s)

4000

5000

Response (%)

0.4

0

40

R-GO/PU (14)

Test A After test B

Re

60

20

200

R-GO/PU (14)

5000

After test B

Time (s)

Response (%)

Normalized Current, I/Io

80

(1) No stretch (2) At 30% stretch (3) Release

(1) 1000

1.2

4000

R-GO/PU (8)

(3)

Test A

60

0

3000

100

Strain (%) Initial (Io ~ 1.5 A)

2000

Time (s)

R-GO/PU (8)

After 1st cyclic test

0.0

0.0

1000

Response (%)

Test A After test B

0.4

1.0

30

100

R-GO/PU (8)

Initial (Io ~ 0.6 A)

Response (%)

Nonrmalized Current, I/Io

1.2 1.0

20

Strain (%)

100

(1)

50

0

(1) No stretch (2) At 30% stretch

(2) 1000

2000

3000

4000

5000

Time (s)

Figure S3.1. Sensing data of the devices with one R-GO/PU fibre layer. (a-b) Current changes of devices (3) and (5) as a function of elongation strain (ε = 0‒30%), respectively. (c) Time-dependent responses of device (5) towards 2.5-ppm NO2 gas before stretch, during ε = 30%, and after release (ε = 0%). (d-f) Strain-dependent current changes and gas responses of device (8) before cyclic stretching test (i.e., test A) and after 5000 stretching cycles at ε = 30% (i.e., test B), respectively. (g-i) Strain-dependent current changes and responses of device (14) before and after test B, respectively.

SI.13

(b) Test A After test B

0.8 0.6

Rel

eas

0.4

R-GO/PU (3\3)

ed

0.2

100

(2) 50

(1) No stretch (2) At 30% stretch

0 0

5

10

15

20

25

30

1000

2000

(e) 150 R-GO/PU (5\3) Test A After test B

0.3 0.2

Rele

ased

0.1 0.0

10

20

30

Strain (%)

4000

50

0

5000

No stretch 1000

2000

40

50

(f) 150

R-GO/PU (5\3) Test A (1)

3000

4000

Time (s) R-GO/PU (5\3)

120 After test B

(2)

90

(4) 60 30 0

After 1st cyclic test 0

120

Response (%)

Normalized Current, I/Io

1.0

Initial (Io ~ 0.5 A)

3000

100

Time (s)

Strain (%)

(d) 1.1

150

(1)

150

After 1st cyclic test

0.0

R-GO/PU (3\3) After test B

Test A

(1) No stretch (2) At 30% stretch (3) At 50% stretch (4) Release

(3) 1000

2000

3000

Time (s)

4000

5000

Response (%)

1.0

(c) 200

200

R-GO/PU (3\3)

Initial (Io ~ 30 nA)

Response(%)

1.2

Response(%)

Normalized Current, I/Io

(a)

90 60 30 0

(1) 1000

(1) No stretch (2) At 30% stretch

(2) 2000

3000

4000

5000

Time (s)

Figure S3.2. Sensing data of the devices with two R-GO/PU fibre layers. (a) Changes in current level of device (3\3) as a function of elongation strain (ε = 0‒30%). (b-c) Time-dependent responses of the device (3\3) towards 2.5-ppm NO2 gas before (test A) and after cyclic stretching test B, respectively. (d-f) Straindependent current changes and gas responses of device (5\3) before and after test B, respectively.

SI.14

0.8

Re le

ase

0.3 0.2

R-GO/PU (5\3\1)

d

(1)

200

(4) (1)

(2)

150

100

(3)

0

0.0 10

20

30

50

40

1000

Strain (%)

Test A (1) After test B

0.6 0.4

Rele

ase

(1)

100

(2)

50

(1) No stretch (2) At 30% stretch 1000

5000

d

90

60

3000

4000

5000

120

R-GO/PU (5\5\3) After test B

Test A (2)

90

(4) (3)

30

0

2000

Time (s)

R-GO/PU (5\5\3)

R-GO/PU (5\5\3)

0.8

0.2

(1) 150

0

120

Iniital (Io ~ 1.5 A)

Response (%)

1.0

4000

3000

2000

After test B

Time (s)

1.2

Normalized Current, I/Io

(1) No stretch (2) At 30% stretch (3) At 50% stretch (4) Release

50

0.1

0

R-GO/PU (5\3\1)

200 Test A

(1) No stretch (2) At 30% stretch (3) At 50% stretch (4) Release

(1)

Response (%)

Normalized Current, I/Io

Test A (1) After test B

0.4

250

250

R-GO/PU (5\3\1)

Response (%)

Initial (Io ~ 0.75 A)

1.0

Response (%)

1.2

60

(1)

30

(1) No stretch (2) At 30% stretch

(2) 0

0.0 0

10

20

30

Strain (%)

40

50

1000

2000

3000

Time (s)

4000

5000

1000

2000

3000

4000

5000

Time (s)

Figure S3.3. Sensing data of the devices with three R-GO/PU fibre layers. (a-c) Strain-dependent current changes and gas responses of device (5\3\1) before and after cyclic stretching test B, respectively. (d-f) Strain-dependent current changes and gas responses of device (5\5\3) before and after test B, respectively.

SI.15

0.4

Rele a

0.3 0.2

se

(1)

200

200 After test B

Test A (1) (2)

150

(4) 100

(1) No stretch (2) At 30% stretch (3) At 50% stretch (4) Release

50

0.1

(3) 0.0

0

(2) 0

10

20

30

40

50

1000

2000

Strain (%)

(2) (1)

100

50

(1) No stretch (2) At 30% stretch 1000

2000

3000

4000

Normalized Current, I/Io

Response (%)

200 After test C

0

1.0

0.3

Test A (1) After test B (2) After test C Re

(1)

lea

se d

0.1 0.0

5000

R-GO/PU (5\3\3\3) Test A (2) 200

(3)

10

20

30

40

(4)

(1)

150 100

(1) No stretch (2) At 30% stretch (3) At 50% stretch (4) Release

50 0

(2) 0

50

1000

2000

3000

4000

5000

Time (s)

R-GO/PU (5\3\3\3) 250

After test B

Response (%)

(2)

150 100 50

(1) No stretch (2) At 30% stretch

0 1000

2000

3000

4000

After test C

200

(2) 150 100 50

(1) (1) No stretch (2) At 30% stretch

0

5000

1000

Time (s) R-GO/PU (8\3\3\1) Test A (1) After test B (2) After test C

0.5

Re

0.4

lea

se

(1)

d

0.2

200

20

30

Strain (%)

5000

150

R-GO/PU (8\3\3\1)

200

After tests B and C (4) 150 (3)

(1) (2)

100

50

40

50

50

0 1000

2000

3000

Time (s)

4000

5000

R-GO/PU (8\3\3\1)

100

(1) No stretch (2) At 30% stretch

0

(2) 10

4000

Test A

0.1

0

3000

Time (s)

Response (%)

Initial (Io ~ 1.1 A)

2000

Response (%)

Response (%)

4000

300

(1)

Normalized Current, I/Io

3000

250

Strain (%)

200

0.0

2000

Time (s)

0.2

R-GO/PU (5\3\3\3)

0.3

(1) No stretch (2) At 30% stretch 1000

5000

R-GO/PU (5\3\3\3)

0.4

5000

300

1.0

(1)

50

300

Initial (Io ~ 1.4 A)

Time (s)

1.2

100

0

1.2

R-GO/PU (5\3\3\1)

250

4000

(2) 150

Time (s)

250

150

3000

Response (%)

Test A (1) After test B (2) After test C

R-GO/PU (5\3\3\1)

R-GO/PU (5\3\3\1)

R-GO/PU (5\3\3\1)

Response (%)

1.0

250

250

Initial (Io ~ 0.8 A)

Response (%)

Normailized Current, I/Io

1.2

(1) No stretch after test B (2) No stretch after test C (3) At 30% stretch (4) At 50% stretch

(1) (2) 1000

2000

3000

4000

5000

Time (s)

SI.16

1.2

250

Test A (1) After test B (2) After test C

0.8 0.6 (1)

Re

lea

se

0.4

d

200

After tests B and C

200

(2) (1) 150

100

50

0.2

(2)

R-GO/PU (8\3\3\3)

Test A

Response (%)

1.0

250

R-GO/PU (8\3\3\3)

R-GO/PU (8\3\3\3) Response (%)

Normalized Current, I/Io

Inital (Io ~ 2A)

(1) No stretch (2) At 30% stretch

0

150 (2)

(3) (1)

100 (4) 50

(1) No stretch after test B (2) No stretch after test C (3) At 30% stretch (4) At 50% stretch

0

0.0 0

10

20

30

Strain (%)

40

50

1000

2000

3000

Time (s)

4000

5000

0

1000

2000

3000

4000

5000

Time (s)

Figure S3.4. Sensing data of the devices with four R-GO/PU layers. (a) Changes in current level of device (5\3\3\1) as a function of elongation strain before and after the 1st cyclic stretching test (i.e., test B, 5000 cycles with ε = 30%) and then 2nd cyclic stretching test (i.e., test C, 5000 cycles with ε = 50%). (b-d) Timedependent gas responses of device (5\3\3\1) before and after two cyclic stretching tests B and C, respectively. (e-h) Strain-dependent current changes and time-dependent gas responses of device (5\3\3\3) before and after tests B and C, respectively. (i-k) Strain-dependent current changes and gas responses of device (8\3\3\1) before and after tests B and C, respectively. (l-n) Strain-dependent current changes and gas responses of device (8\3\3\3) before and after tests B and C, respectively.

SI.17

0.4

Re

(1) 0.3

lea sed

0.2

150

150

(1) 100

(2) 50

(1) No stretch (2) At 30% stretch (3) At 50% stretch

0.1 0

(2) 10

20

30

40

1000

50

0.4

Re

lea

sed

(1)

150

Response (%)

Test A (1) After test B (2) After test C

0.6

20

0

5000

1000

30

40

50

2000

3000

4000

5000

Time (s) R-GO/PU (8\5\3\3\1) After tests B and C 150

(2) 100

(1)

50

0

10

(1) No stretch after test B (2) No stretch after test C (3) At 30% stretch (4) At 50% stretch

Test A

(1) No stretch (2) At 30% stretch (3) At 50% stretch

(3)

(2) 0

50

200

0.2 0.0

4000

(4)

100

R-GO/PU (8\5\3\3\1)

R-GO/PU (8\5\3\3\1)

0.8

3000

(3)

(2)

0

200

1.2

Inital (Io ~ 1.5 A)

2000

(1)

Time (s)

Strain (%) 1.0

(3)

Response (%)

0.0

0

Normalized Current, I/Io

After tests B and C

Test A

Response (%)

Test A (1) After test B (2) After test C

R-GO/PU (8\3\3\3\3)

R-GO/PU (8\3\3\3\3)

Initial (Io ~ 1.9A) R-GO/PU (8\3\3\3\3)

Response (%)

Normalized Current, I/Io

1.0

200

200

1.2

1000

Strain (%)

2000

3000

Time (s)

4000

5000

(1)

(2)

100

50

(1) No stretch after test B (2) No stretch after test C (3) At 30% stretch (4) At 50% stretch

(3) (4)

0 0

1000

2000

3000

4000

5000

Time (s)

Figure S3.5. Sensing data of the devices with five R-GO/PU layers. (a-c) Strain-dependent current changes and time-dependent gas responses of device (8\3\3\3\3) before and after cyclic stretching tests B and C, respectively. (d-f) Strain-dependent current changes and gas responses of the device (8\5\3\3\1) before and after cyclic stretching tests B and C, respectively.

SI.18

Table S3.1. Comparison of R-GO/PU devices under stretching tests of A, B, and C. R-GO/PU device

Sustainability ((%))

Gas responsivity ((ΔI/I(%)) / (%))

Remarks

A

B

C

A

B

C

(3)

30











Low conductance, very high noise

(5)

30





35/0 30/30





The sensing at εA = 30%.

(8)

50

0



60/0 60/30

45/0



Did not pass test B. The sensing signal was degraded at εB = 0%.

(14)

50

30



120/0 105/30

120/0 120/30



Passed test B. The sensing signal was degraded at εB = 30%.

(3\3)

30

0



130/0 120/30

150/0



Did not pass test B. The sensing signal was degraded at εB = 0%.

(5\3)

50

30



110/0 120/30 115/50

120/0 120/30



Passed test B. The sensing signal was degraded at εB = 30%.

(5\3\1)

50

30



160/0 175/30 165/50

150/0 145/30



Passed test B. The sensing signal was degraded at εB = 30%.

(5\5\3)

50

30





Passed test B. The sensing signal remained well at εB = 30%.

(5\3\3\1)

50

30

30

190/0 175/30 180/50

180/0 185/0

210/0 210/30

Passed test C. The sensing signal remained at εC = 30%.

(5\3\3\3)

50

30

30

200/0 200/30 200/50

210/0 210/30

160/0 210/30

Passed test C. The sensing signal remained at εC = 30%.

(8\3\3\1)

30

30

50

140/0 150/30

160/0

145/0 160/30 150/50

Passed test C. The sensing signal remained at εC = 30% but got degraded at εC = 50%.

(8\3\3\3)

30

30

50

160/0 150/30 140/50

150/0

150/0 140/30 145/50

Passed test C. The sensing signal remained at εC = 50%.

(8\3\3\3\1)

50

50

50

110/0 120/30 125/50

120/0 125/30 130/50

130/0 130/30 125/50

Passed test C. The sensing signal remained well at εC = 50%.

(8\3\3\3\3)

50

50

50

145/0 135/30 125/50

150/0

135/0 150/30 150/50

Passed test C. The sensing signal remained well at εC = 50%.

(8\5\3\3\1)

50

50

50

125/0 145/30 150/50

120/0

120/0 125/30 120/50

Passed test C. The sensing signal remained well at εC = 50%.

90/0 75/30 70/0 70/30 70/50

signal

was

degraded

SI.19

Supporting Information 4 1. Sensing performance of devices towards various NO2 concentrations

0.8

40

0.4

30

0.2 35

0.05

15

0.1

No stretch

0

30 2000

4000

6000

8000

10000

120

Sensing NO2

1.6

0.04

60 40

0.05

0.03

0.1

0.2 20

No stretch after test B 3000

6000

9000

12000

0

15000

Time (s) 0.06

R-GO/PU (5\3\3\1) Sensing NO2

250

2.5 ppm 200

0.05

Current (A)

80

0.4

Time (s)

(c)

100

0.8

0.02

12000

140

2.5 ppm

R-GO/PU (14) 0.05

Response (%)

Current (A)

Sensing NO2

(b)

45

2.5 ppm

1.6 150

0.04

0.03

0.02

0.8 100

0.05 0.4 0.2 0.1

50

95% recovery

~ 3 hr

No stretch after test C 1000

5000

9000

13000

Response (%)

1.6

Current (A)

R-GO only

Response (%)

(a) 45

0

17000

Time (s)

Figure S4.1. Time-dependent gas sensing measurements at room temperature. (a) Sensing responses of R-GO film device to NO2 gas diluted in dry air with a concentration range of 50 ppb to 2.5 ppm. (b-c) Sensing responses of the R-GO/PU fibre device (14) after test B and the R-GO/PU fibre device (5\3\3\1) after test C towards various NO2 concentrations in dry air, respectively.

SI.20

2. Calculation of detection limit Limit of detection (LOD) is defined as the lowest gas concentration to which the device can respond with a signal magnitude about 3 times higher than the signal noise (IUPAC). Here are the steps to calculate LOD which was described in the previous studies.[2,3] 1) Determine noise level. In general, the noise level of a device can be calculated via the root-mean squared deviation (RMS). After choosing 11 data points (n = 11) at the baseline, the data was plotted as response (Y, %) versus time (X, sec) and then a polynomial fit (5th order) was executed in the OriginPro software (OriginLab). After taking regular residual (Y - Ῡ) from the polynomial fit, the RMS noise (= √Σ(Y − Ῡ)2 / (𝑛 − 1)) is determined as shown in Tables S4.1 and S4.2. 2) Determine gas sensitivity in linear response region. Data extracted from time-dependent sensing measurements at various gas concentrations were re-plotted as sensing response (%) versus gas concentration (ppm). Next, linear fitting was performed for the linear region of the response-toconcentration plot to extract the slope (i.e. sensitivity) by using OriginPro, shown in Figure S4.2. 3) Calculate LOD by using the below equation (*). The results are shown in Table S4.3. ____________________________________________________________________________ (*) Relationship of response (y, %) and gas concentration (x, ppm) in a linear response region 𝑦 = 𝑎+𝑏×𝑥 ↔ ↔

𝑥=

(𝑦−𝑎) 𝑏 3 × 𝑛𝑜𝑖𝑠𝑒 (%)

𝐿𝑂𝐷 (𝑝𝑝𝑚) = 𝑠𝑒𝑛𝑠𝑖𝑡𝑖𝑣𝑖𝑡𝑦 (%/𝑝𝑝𝑚)

SI.21

Table S4.1. Calculation of root-mean-square noise of the R-GO device (a) and the R-GO/PU fibre devices (14) and (5\3\3\1) (b and c, respectively)

Time (s)

(a) R-GO only Signal Y-Ῡ (%) (%)

420 430 440 450 460 470 480 490 500 510 520

0.17793 0.16031 0.14792 0.13106 0.11302 0.09925 0.08019 0.06661 0.04986 0.0378 0.0212

(Y-Ῡ)^2

Time (s)

2.35E-04 -1.04E-03 1.45E-03 -3.10E-05 -1.68E-03 1.51E-03 -8.62E-04 1.18E-03 -1.24E-03 5.92E-04 -1.03E-04

5.53E-08 1.08E-06 2.10E-06 9.60E-10 2.83E-06 2.29E-06 7.44E-07 1.39E-06 1.55E-06 3.51E-07 1.06E-08

3040 3050 3060 3070 3080 3090 3100 3110 3120 3130 3140

RMS Noise

1.11E-03

(b) R-GO/PU (14) after test B Signal Y-Ῡ (Y-Ῡ)^2 (%) (%) 0.05248 0.08918 0.1646 0.21576 0.17054 0.04514 0.08652 0.1361 0.04964 0.04485 0.15677

1.39E-02 -3.26E-02 2.37E-03 4.63E-02 1.74E-02 -7.83E-02 -3.89E-03 7.13E-02 -7.61E-03 -3.41E-02 1.54E-02

1.94E-04 1.06E-03 5.63E-06 2.14E-03 3.02E-04 6.12E-03 1.51E-05 5.09E-03 5.79E-05 1.17E-03 2.37E-04

RMS Noise

4.05E-02

(c) R-GO/PU (5\3\3\1) after test C Time Signal Y-Ῡ (Y-Ῡ)^2 (s) (%) (%) 480 490 500 510 520 530 540 550 560 570 580

0.07994 0.18092 0.06924 0.12374 0.1238 0.08907 0.16201 0.0836 0.07494 0.18632 0.06919

-1.47E-02 5.50E-02 -6.00E-02 4.25E-03 1.56E-02 -1.39E-02 5.54E-02 -3.36E-02 -5.24E-02 6.24E-02 -1.81E-02

2.17E-04 3.02E-03 3.60E-03 1.81E-05 2.42E-04 1.92E-04 3.07E-03 1.13E-03 2.75E-03 3.89E-03 3.27E-04

RMS Noise

4.30E-02

SI.22

Table S4.2. Calculation of root-mean-square noise of the R-GO/PU fibre device (8\3\3\3\1) before stretching (a), after test C (b), and at 30% and 50% elongation (c and d, respectively).

Time (s)

(a) (8\3\3\3\1) before stretch Signal Y-Ῡ (Y-Ῡ)^2 (%) (%)

Time (s)

(b) (8\3\3\3\1) after test C Signal Y-Ῡ (Y-Ῡ)^2 (%) (%)

560 570 580 590 600 610 620 630 640 650 660

-0.03002 0.02417 0.0453 0.01983 0.12929 0.07619 0.1074 0.10127 0.16473 0.23013 0.20976

550 560 570 580 590 600 610 620 630 640 650

0.0692 0.04832 0.07455 0.02719 0.00196 0.08041 0.1358 0.1194 0.18274 0.15375 0.08196

3.61E-03 -2.58E-04 -1.45E-02 -5.52E-03 3.93E-02 -2.01E-02 4.65E-03 -1.09E-02 -1.57E-02 3.09E-02 -1.15E-02

1.31E-05 6.65E-08 2.10E-04 3.05E-05 1.55E-03 4.06E-04 2.16E-05 1.19E-04 2.47E-04 9.57E-04 1.32E-04

RMS Noise

1.92E-02

-3.32E-04 -1.10E-02 3.12E-02 -7.34E-03 -3.91E-02 1.52E-02 3.24E-02 -2.59E-02 9.58E-03 -8.25E-03 3.55E-03

1.10E-07 1.22E-04 9.76E-04 5.38E-05 1.53E-03 2.30E-04 1.05E-03 6.68E-04 9.18E-05 6.80E-05 1.26E-05

RMS Noise

2.19E-02

Time (s)

(c) (8\3\3\3\1) at εC = 30 % Signal Y-Ῡ (Y-Ῡ)^2 (%) (%)

Time (s)

(d) (8\3\3\3\1) at εC = 50 % Signal Y-Ῡ (Y-Ῡ)^2 (%) (%)

1000 1010 1020 1030 1040 1050 1060 1070 1080 1090 1100

0.05715 0.0762 0.01787 0.1749 0.16258 0.1945 0.21464 0.27229 0.29577 0.23419 0.2511

650 660 670 680 690 700 710 720 730 740 750

0.1288 0.04313 0.28663 0.18841 0.27166 0.33905 0.03939 0.19768 0.29577 0.12977 0.10872

-2.26E-03 2.19E-02 -6.00E-02 5.73E-02 -5.73E-04 -1.14E-02 -2.51E-02 1.16E-02 2.85E-02 -2.65E-02 6.51E-03

5.11E-06 4.80E-04 3.60E-03 3.28E-03 3.28E-07 1.30E-04 6.30E-04 1.35E-04 8.12E-04 7.00E-04 4.24E-05

RMS Noise

3.13E-02

4.34E-02 4.34E-02 4.34E-02 4.34E-02 4.34E-02 4.34E-02 4.34E-02 4.34E-02 4.34E-02 4.34E-02 4.34E-02

1.89E-03 1.89E-03 1.89E-03 1.89E-03 1.89E-03 1.89E-03 1.89E-03 1.89E-03 1.89E-03 1.89E-03 1.89E-03

RMS Noise

4.55E-02

SI.23

140

30 25 Equation

20

y = a + b*x

Instrumental Weight Residual 23.7994 Sum of Squares 0.98513 Pearson's r 0.95573 Adj. R-Squ

15

R-GO/PU (14)

120

Without stretch

10 B

After test C

100 80 60 40 20

Value

Equation

y = a + b*x

Weight Residual Sum of Squares Pearson's r

Instrumental

Adj. R-Squar

15.63181 0.99687

4.5196

0.91633

Slope

60.829

7.50096

Value

0

Intercept

C

150

100

Equation

Slope

50

Value

Standard Err

-0.32395

0.35834

83.7313

3.83572

y = a + b*x

No Weighting Weight Residual 77.09031 Sum of Squares 0.995 Pearson's r 0.9867 Adj. R-Squa

0.99166

Standard Er

Intercept

R-GO/PU (5\3\3\1)

200

After test B

Response (%)

Response (%)

35

RGO only

Response (%)

40

0

Intercept

D

Slope

Standard Er

1.78542

3.43186

143.3954

8.31128

2.0

2.5

5 0.0

0.5

1.0

1.5

2.0

2.5

0.0

NO2 concentration (ppm)

0.5

1.0

1.5

2.0

2.5

0.0

0.5

NO2 concentration (ppm)

1.0

1.5

NO2 concentration (ppm)

160 140

140

R-GO/PU (8\3\3\3\1) Before stretch in test A

120

R-GO/PU (8\3\3\3\1) After test C

Response (%)

Response (%)

120 100 80 60 40

Equation

y = a + b*x

Weight Residual Sum of Squares Pearson's r

No Weighting 80.44792 0.9852 0.95593

Adj. R-Squar

Value

20 E

Intercept

8.7182

Slope

96.151

1.5

2.0

100 80 60

Equation

y = a + b*x

Weight Residual Sum of Squares Pearson's r

No Weighting

40 20

Adj. R-Squar

4.11924 0.99895 0.99721 Value

Standard Err 5.45259

F

0

11.82833

Standard Err

Intercept

-1.2712

0.7933

Slope

72.6104

1.92122

2.0

2.5

0 0.0

0.5

1.0

2.5

0.0

NO2 concentration (ppm)

0.5

1.0

1.5

NO2 concentration (ppm) 160

160

R-GO/PU (8\3\3\3\1)

140

At 30% stretch after test C

140 120

Response (%)

120

Response (%)

R-GO/PU (8\3\3\3\1) At 50% stretch after test C

100 80 60 40

Equation

y = a + b*x

Weight Residual Sum of Squares Pearson's r

Instrumental

Adj. R-Squa

0.98351

30.1512 0.9938

20

Value G

0 0.0

0.5

1.0

Intercept Slope

1.5

0.73583

Standard Err

80 60 40

8.29167

2.0

2.5

Equation

y = a + b*x

Weight Residual Sum of Squares Pearson's r

Instrumental

Adj. R-Squar

20

15.47674 0.99654 0.98964 Value

0.82893

128.3492

NO2 concentration (ppm)

100

H

0 0.0

0.5

1.0

Standard Err

Intercept

-7.13382

0.97039

Slope

127.7397

7.53322

2.0

2.5

1.5

NO2 concentration (ppm)

Figure S4.2. Determination of gas sensitivity of the devices based on R-GO only and R-GO/PU fibres. (a) A plot of response (%) versus NO2 concentration (ppm) and linear fitted data (blue line) of the R-GO device. (b-c) NO2 concentration-dependent response plots and linear fitted data of the R-GO/PU device (14) after test B and the R-GO/PU fibre device (5\3\3\1) after test C, respectively. (d-g) NO2 concentration-dependent response plots and linear fitted data of the R-GO/PU fibre device (8\3\3\3\1) before stretch, after test C, and at static elongation strains (εC = 30 and 50%), respectively.

SI.24

Table S4.3. Calculation of detection limit values of R-GO and R-GO/PU fibre devices

Samples

RMS Noise

b (= slope) Standard Error

x = LOD (ppm)

R-GO only before stretch

1.11E-03

60.8291

7.50096

5.47E-05 ± 7.70E-06

R-GO/PU (14) after B test

4.05E-02

83.73136

3.83572

1.45E-03 ± 6.97E-05

(5\3\3\1) after C test

4.30E-02

143.395

8.31128

9.00E-04 ± 5.54E-05

(8\3\3\3\1) before stretch

1.92E-02

96.151

11.82833

5.99E-04 ± 8.40E-05

(8\3\3\3\1) after C test

2.19E-02

72.61048

1.92122

9.05E-04 ± 2.46E-05

(8\3\3\3\1) at 30% stretch

3.13E-02

128.349

8.29167

7.32E-04 ± 5.06E-05

(8\3\3\3\1) at 50% stretch

4.55E-02

127.7397

7.53322

1.07E-03 ± 6.70E-05

SI.25

Supporting Information 5 The information of Supplementary Video SV The video clip SV was obtained by using a portable webcam (2.0 megapixels). To reduce the file size, the playing speed is set to play 4 times faster (4x) than normal speed. The sensing setup was shown in Figure S5.1 to observe and compare dynamic responses towards strain (finger motions) and gas exposures (40 s / 1-ppm NO2 pulse) at ambient conditions (23oC, 30% RH). Because gas concentrations can quickly be diluted by ambient air, the total air flow (humidified to about 30% RH) was increased from 1000 to 3000 sccm in order to maintain the concentration at the gas outlet. And for safety, the gas exposures were totally less than 1.5 min and a fume hood (tube) was opened during the measurement. As shown in Figure S5.1(d), the current intensity (amplitude) of the R-GO/PU fibre device (8\3\3\3\1) was reduced when the finger was bent, and it then returned to the baseline when the finger was straightened. The strain responses (ε ≈ 50%) were about I/Io ≈ 0.1. Regarding gas sensing, the device responded well to 1-ppm NO2 (ΔI/I ≈ 130% to 160% after 40-sec exposures). Here, the gas responses in the practical measurement are higher than the previous data of the R-GO/PU fibre device (8\3\3\3\1) as shown in Figure S5.2 (linear fit of response-to-concentration plots) where the responses to 1-ppm NO2 gas (i.e., slope) were about 70‒120% after 2-min exposure. This reveals that there can be many factors from ambient environment which can affect the sensing signal in addition to strain. To understand what enhanced the gas sensing responses, additional sensing measurements were carried out as shown in Figure S5.2. First, Figures S5.2(a-b) display humidity effects on the current level and sensing capability at 23oC, respectively. This is because H2O molecules have a strong polarization which can disturb local electric field and charge transfer on the R-GO sensing network.[4] Both the current level and gas response decreased when the relative humidity (RH) level increased from dry air to 30% RH probably because additional interactions between H2O and NO2 (acid vapour formation) that were in addition to the interactions between R-GO and NO2. This indicates that humidity did not help to enhance NO2 response.

SI.26

As well known, temperature also exerted some effects on sensing. In our opinion, heat could not only transfer from the gloved finger to the attached sample but also be generated during measurement, resulting in a relative temperature of the device (estimated 30oC) higher than room temperature (~ 23oC). Therefore, next measurements were performed to check temperature effects and shown in Figures S5.2(c,d). Figure S5.2(c) reveals a slight increase of gas response in dry air when the chamber temperature was increased from 23 to 30oC, probably because of more charges thermally induced to react with NO2 molecules. More interestingly, a significant enhancement in gas response was observed in Figure S5.2(d) when the device had been stabilized in 30oC and 30% RH. The gas responses (ΔI/I ≈ 120%‒160% after 40-s exposure) in Figure S5.2(d) seem to match the data in the practical measurement (Figure S5.1d). Besides, the mass flow as well as pressure may affect the number of interactions between gas molecules and the sensing channel. To understand about it, we measured changes in current level and gas response towards increasing the mass flow rate from 0 to 3000 sccm as shown in Figure S5.2(e,f). Figure S5.2e shows a current decrease upon increasing the mass flow rate. Although it is negligible in comparison to the NO2 response (Figure S5.2(e)), a small enhancement of NO2 response in the 3000-sccm flow was also observed because of the increased number of gas molecules in the chamber in comparison to that in the 1000-sccm flow (Figure S5.2(f)). Through these observations, we may assume that enhancement of gas responses in the practical measurements could come from the increase of heat (from finger motions and increased resistance during a long measurement), increased mass flow rate, and water removal (humidity reduction by the air + gas flow and the heat). In summary, the practical measurements did not show a significant signal ratio of gas response/strain, indicating the device performance still needs more improvement for practical use in wearable applications. Besides, there are many factors that can interfere with the sensing signal, resulting in strong requirements of further studies to determine and solve the effects of the surrounding environments for reliable chemical sensing applications.

SI.27

Figure S5.1. Illustration of gas sensing measurement using the R-GO/PU fibre device (8\3\3\3\1). (a) Setup for sensing measurement in ambient conditions (23oC, 30% RH). (b-c) Pictures of the fibre device (8\3\3\3\1) placed on a finger when straighten (channel size = 0.7 cm) and bent (actual channel size = curve length = 1‒1.05 cm), respectively. Here, with the finger bent, the device was not only bent but also stretched with a strain of about 50%. (d) A frame in Supporting Video showing the response of the device towards strain and 1-ppm NO2 gas exposure (about 40 s/pulse).

SI.28

(a) 0.50

(b) 150

R-GO/PU (8\3\3\3\1)

R-GO/PU (8\3\3\3\1) Sensing at different humidity levels

120

Response (%)

0.00

Dry air 10% RH 20% RH 30% RH

-0.25

(1) (2)

(3)

90

(4) 60 30 0

(1) Dry air (2) 10% RH

(3) 20% RH (4) 30% RH

2000

4000

-0.50 -6

-4

-2

0

2

4

6

1000

3000

Voltage (V)

(c)

Time (s)

(d)

150

R-GO/PU (8\3\3\3\1)

R-GO/PU (8\3\3\3\1)

300 o

90

Current (A)

Response (%)

0.28 0.24 Sensing at 30% RH and 30 C

120 Sensing at different temperature

(1)

5000

(2)

60

30

NO2

250

NO2

0.20

200 40 s ~ 160 %

0.16 40 s of exposure I/I ~ 120 %

0.12

150 100 50

(1) At 23oC in dry air (2) At 30oC in dry air

0

Response ()

Current (A)

0.25

0.08 Dry, 30oC

0

30% RH, 30oC

0.04 1000

2000

3000

4000

5000

0

200 500

Time (s)

Time (s)

(f)

R-GO/PU (8\3\3\3\1)

1.5 no air flow (ambient pressure) 1.4

1000 sccm dry air

1.3

50 0.5 ppm NO2 in the 1000-sccm flow 40

2000 sccm dry air

0.5 ppm NO2 3000 sccm dry air

3000 sccm dry air

1.2

Response (%)

Current (A)

(e) 1.6

30 20 0.5 ppm NO2 in the 3000-sccm flow

10 0

1.1 200

400

1000 1500 2000 2500 3000 3500

600

Time (s)

800

1000

R-GO/PU (8\3\3\3\1) 0

100

200

300

400

500

Time (s)

Figure S5.2. Sensing measurements at different conditions. (a-b) I-V characteristics and responses of device (8\3\3\3\1) to 1-ppm NO2 (2 min-exposures) in dry air and humidified air (10 to 30% RH), respectively. (c-d) Responses to 1-ppm NO2 (2-min exposures) at 23 to 30oC in dry air and at 30oC in 30% RH, respectively. (e) Current changes towards varied mass flow conditions and NO2 (diluted in a 3000sccm flow). (f) Responses to 0.5 ppm NO2 diluted in 1000 - 3000 sccm of the total flows, respectively.

SI.29

REFERENCES [1] 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 [2] 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 [3] 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-Organic Vapour Sensor Using Inkjet-Printed Reduced Graphene Oxide Angew. Chem. Int. Ed. 49 2154–2157 [4] 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 4329–4338 -----------------------------------------------------

SI.30