Soft and flexible Interpenetrating Polymer Networks hosting ...

Solar Energy Materials & Solar Cells 127 (2014) 33–42

Contents lists available at ScienceDirect

Solar Energy Materials & Solar Cells journal homepage: www.elsevier.com/locate/solmat

Soft and flexible Interpenetrating Polymer Networks hosting electroreflective poly(3,4-ethylenedioxythiophene) Laurent J. Goujon a, Pierre-Henri Aubert a,n, Laurent Sauques b, Frédéric Vidal a, Dominique Teyssié a, Claude Chevrot a a Laboratoire de Physicochimie des Polymères et des Interfaces (EA 2528), Fédération “Institut des Matériaux”, Université de Cergy-Pontoise, 5 mail Gay-Lussac, 95031 Cergy-Pontoise Cedex, France b Ministère de la Défense/Direction Générale pour l'Armement, 7/9 rue des Mathurins, 92221 Bagneux Cedex, France

art ic l e i nf o

a b s t r a c t

Article history: Received 20 December 2013 Received in revised form 13 March 2014 Accepted 14 March 2014

In this paper, we designed a flexible electroreflective device (ERD) with tunable optical properties in the near and mid infrared spectra. The ERD is made of a poly(3,4-ethylenedioxythiophene) (PEDOT) layer interpenetrated in a soft and flexible nitrile butadiene rubber (NBR)/polyethylene oxide (PEO) interpenetrating polymer network (IPN). The composition, structure and infrared properties of the semi-IPN were optimized according to the polymerization time of the EDOT monomer within the IPN. After the addition of an ionic liquid (IL), namely N-ethylmethylimidazolium bis(trifluoromethanesulfonyl)imide (EMITFSI), assuming the ionic conductivity of the system, the device exhibits middle infrared and tunable optical properties between 2 mm and 20 mm upon electrical switching between 1.5 V and þ1.5 V. The modulation of the reflectivity in bands II (3–5 mm) and III (8–12 mm) regions is 26% and 41% respectively. Due to the presence of the soft NBR network, both IPN and ERD reveal improved mechanical properties compare to the single network and the ERD-only based on PEO. For instance, ERDs made with PEO/PEDOT semi-IPN possesses an elongation at break below 10%, for a stress at break lower than 1.4 MPa, while ERDs containing NBR show elongations at break from 45% to 15% for stresses at break from 3 to 8 MPa, depending on the amount of incorporated PEDOT. In parallel, the NBR/PEO/ PEDOT devices showed robust and flexible behavior. & 2014 Elsevier B.V. All rights reserved.

Keywords: Electroreflective device Interpenetrating Polymer Network Nitrile butadiene rubber Poly(3,4-ethylenedioxythiophene) Poly(ethylene oxide)

1. Introduction Electronic conducting polymers (ECP) are very attractive materials because of their promising electronic, optical and electrochemical properties [1]. Among the numerous applications that include ECPs, the optical modulation in the visible and infrared region opens up interesting opportunities for developing electrochromic devices (ECD) [2–5] or electroreflective devices (ERD), both in the visible [6,7] and infrared ranges [8–10]. However, usually the devices are designed on multilayer architectures that impair the sustainability of such applications. Our research group has recently developed devices based on semi-Interpenetrating Polymer Networks (semi-IPN), in which the architecture is simpler than that of multilayered devices [6,8,9,11,12]. Typically, the ECP – a linear polymer – is interpenetrated in a supporting polymer electrolyte like polyethylene oxide. This architecture combines the properties of each polymer network depending on their nature

n

Corresponding author. Tel.: þ 33 134257088; fax: þ 33 134257000. E-mail address: [email protected] (P.-H. Aubert).

http://dx.doi.org/10.1016/j.solmat.2014.03.036 0927-0248/& 2014 Elsevier B.V. All rights reserved.

and affords morphological stability compared to conventional polymer blends [13]. In the case of electro-optical devices, adequate combinations of the optical and electronic properties of ECP with the ionic conductivity properties of a polymer network at least give rise to a promising class of materials. This was recently demonstrated in our group with poly(3,4-ethylenedioxythiophene) (PEDOT) [14] or poly(3,4-(2,20 -dimethylpropylene)dioxythiophene) (PProDOT-Me2) [15] as ECP and poly(ethylene oxide) (PEO) as the ionic conducting network. The potential resulting application depends on the PEDOT content, i.e., if the PEDOT weight content is low (o0.3 wt%), the PEO/PEDOT semi-IPN is transparent in the visible range and can be used to design electrochromic devices [14]; while if the weight content is higher, the semi-IPN becomes opaque and the PEDOT concentration at the surface is sufficient to obtain infrared electroreflective properties [16]. These results demonstrate the adaptability of the self-supported semi-IPNs in electro-optical applications. Despite the good optical properties obtained both in the visible and middle infrared regions (transmission variation up to 33% in visible at 630 nm [14] and reflectivity variation up to 40% in infrared at 25 mm [16]) such semi-IPNs usually suffer from poor mechanical properties: e.g., with a PEO/PEDOT semi-IPN in which

34

L.J. Goujon et al. / Solar Energy Materials & Solar Cells 127 (2014) 33–42

the PEDOT content is 1 wt% and swollen with 8 wt% EMITFSI as electrolyte, the elongation at break is below 10% for a stress at break lower than 1.4 MPa at room temperature. Therefore, such a material is drastically brittle. In order to develop more resilient electro-optical devices, we fabricated a polymeric matrix based on IPN architecture. This matrix, whose elaboration and characterizations were described previously [17] combines the good ionic conductivity of a PEO network (in the presence of EMITFSI) and the satisfactory thermo-mechanical properties of a nitrile butadiene rubber (NBR) network containing ionic liquids. Indeed the presence of high molecular weight NBR in matrices should ensure convenient mechanical properties such as elasticity and flexibility and makes those matrices appropriated in view of their future integration in complex systems [17]. In this paper, we report the elaboration and the characterization of soft and flexible ERDs containing on both faces a thin layer of PEDOT interpenetrated in a NBR/PEO IPN. Typically, the ECP is obtained by a chemical oxidative polymerization of the conjugated monomer firstly incorporated by swelling in the pre-formed NBR/ PEO matrix. The oxidizing agent for the polymerization of the EDOT is FeCl3, leading to the NBR/PEO/PEDOT architecture where the PEDOT is formed predominantly on both faces of the IPN film. After the addition of an ionic liquid, such a monoblock architecture exhibits infrared modulation properties upon the application of a low bias voltage ( 71.5 V) between the two faces. The optical and mechanical properties were optimized based on PEDOT and IL contents. One significant point is that the devices are able to operate in open-air conditions and exhibit electrochemical and optical switches on the active surface, which is not covered by gold contacts. Furthermore, the presence of the elastomeric network (NBR) considerably improved the mechanical properties of the IPN matrix.

(thickness ¼ 140 mm) were obtained with a 54/46 (wt/wt) ratio of both partners. The NBR/PEO matrix possesses good mechanical properties due to the presence of NBR since the elongation at break is 110% for a stress of 1.2 MPa [17]. Then NBR/PEO films were swollen with 143 wt% (7 3 wt%) EDOT at room temperature and finally immersed in an aqueous FeCl3 solution (1.5 mol L 1) at 40 1C in order to promote the oxidative polymerization of EDOT. The 143 wt% ( 73 wt%) EDOT swelling degree was chosen after a preliminary study on the effect of EDOT swelling degree on the maximum variations of the semi-IPN reflectivity (Δ%R – after chemical oxidation/reduction) and on the amount of incorporated PEDOT within the NBR/PEO IPN, for 1 h EDOT polymerization at 40 1C. The results show that the maximum Δ%R and the PEDOT contents were reached a maximal plateau for an EDOT swelling degree higher than 60% (results not shown). We worked with EDOT swelling degrees around 143 wt% because of an easier control and a better reproducibility. The PEDOT synthesis was then optimized by adjusting the immersion/polymerization time between 10 and 75 min. The films were washed several times with methanol (until methanol wash bath is colorless) to remove the excess FeCl3 and non-reacted EDOT. Then NBR/PEO/PEDOT semiIPNs were dried 3 h at 50 1C under vacuum. Finally, the film edges were cut to obtain a three-layer architecture (active electrode/ electrolyte-polymer network architecture/counter-electrode) with a final thickness of 110 mm. Sulfur elemental analyses were performed on all samples allowing the determination of the PEDOT content. The depth of PEDOT interpenetration within the NBR/PEO IPN, determined by scanning electron microscopy (SEM), varied from 3 to 20 mm according to the EDOT polymerization time. Following the NBR/PEO/PEDOT semi-IPNs synthesis, the semi-IPNs were soaked for different times in EMITFSI, used as an electrolyte, to obtain the final electroreflective device (ERD).

2. Experimental section

2.2. Measurements

Nitrile butadiene rubber (NBR) with 44 wt% acrylonitrile content (Mw ¼230,000 g mol 1, Perbunan 4456F Lanxess), dicumyl peroxide (DCP, Aldrich, 98%), poly(ethylene glycol) dimethacrylate (PEGDM, Mn ¼750 g mol 1, Aldrich), poly(ethylene glycol)methyl ether methacrylate (PEGM, Mn ¼ 475 g mol 1, Aldrich) were used as received. Anhydrous iron(III) chloride FeCl3 (Acros, 99%), nitrosonium tetrafluoroborate (NOBF4) (Acros, 97%), hydrazine (Aldrich, 35 wt% in water), cyclohexanone (Acros, 99.8%), dichloromethane (VWR, 99%), methanol (VWR, 99.9%), acetonitrile (Acros, 99%) and N-ethylmethylimidazolium bis(trifluoromethane) sulfonimide (EMITFSI, Solvionic, electrochemical grade 99%) were used as received. 2,2-Azobis(isobutyronitrile) (AIBN) (Aldrich) was freshly recrystallized from methanol before use. 3,4-Ethylenedioxythiophene (EDOT, Bayer) was distilled under reduced pressure at 70 1C.

The swelling degree in network or IPN samples was defined as SD ¼ W S W 0 =W 0 where W0 and WS are the weight of the sample before and after swelling by a compound (monomer, solvent, ionic liquid). The PEDOT contents (% PEDOT), within the NBR/PEO/PEDOT semi-IPNs, were determined by sulfur elemental analysis, specific to the PEDOT, according to the following calculation: % PEDOT ¼ % SðM EDOT =M Sulf ur Þ, where % S is the weight percent of sulfur, Msulfur, the sulfur molar weight (32 g mol 1) and MUnit, the molar weight of a PEDOT repetition unit (140 g mol 1). The PEDOT distribution within the NBR/PEO matrix was mapped with a SEM model CARLZEISS AG-ULTRA 55 GEMINI, connected to an X-ray analysis system (Energy Disperse X-ray analysis, EDAX), the same model. Electrochemical characterizations of PEDOT within NBR/PEO/ PEDOT films were measured using a potentiostat/galvanostat EG&G Princeton Applied Research Model 273 A. A NBR/PEO/PEDOT film (1.3 cm2), swollen with 40 wt% EMITFSI, was pressed in a stainless steel grid and used as the working-electrode. The counter-electrode and the pseudo-reference electrode were a platinized titanium grid and a silver wire respectively. All potentials are referred to the Ag wire calibrated with a ferrocene redox probe (E1Fc þ =Fc ¼ 0:44 V versus Ag wire). Cyclic voltammetry was performed in EMITFSI at 50 mV s 1 of scan rate between 1.5 V and þ1.5 V under argon atmosphere. The NIR optical reflection was measured using an integrated sphere JASCO ISN-470 mounted onto a JASCO V-570 spectrophotometer. The scan rate was set to 1000 nm min 1 between 0.8 and 2.5 mm. MIR characterizations were performed from 2.5 to 20 mm using a Bruker Equinox 55 spectrophotometer fitted with

2.1. Elaboration of NBR/PEO/PEDOT semi-IPN The host matrix, a free-standing NBR/PEO (54/46 wt/wt) IPN, was prepared through a two-step process as previously described [17]. Briefly, IPNs were synthesized using a sequential process in which the PEO network is interpenetrated in the NBR network. First, the NBR single network was obtained by cross-linking the raw NBR in the presence of DCP (2 wt%) for 30 min at 180 1C. Then the NBR network was swollen with a mixture of PEGDM, PEGM (PEGDM/PEGM 50/50 (wt/wt)) and AIBN in cyclohexanone. The free-radical copolymerization of PEGDM and PEGM (1 h at 70 1C plus 1 h at 100 1C), in the presence of AIBN (3 wt% with respect to the total amount of PEGM and PEGDM), leads to the PEO network formation within the NBR network. After cyclohexanone evaporation at 70 1C under vacuum for 24 h, thin NBR/PEO films


35

an integrated Bruker A562G sphere. The maximum variations of semi-IPN reflectivity (Δ%R) were defined as Δ%R¼%Rmax %Rmin, where %Rmax and %Rmin are the maximum and the minimum reflectivities measured for the active surface in oxidized state and in reduced state, respectively. In order to determine the maximal %Rmax in the oxidation state, without influence of electrochemical yield in the PEDOT oxidation within the ERD, the NBR/PEO/PEDOT semi-IPNs, free of electrolyte, were chemically oxidized for 30 min by soaking the samples in a acetonitrile solution containing nitrosonium tetrafluoroborate (NOBF4, 5 10 2 mol L 1) and dried few minutes under air at room temperature. A similar procedure was employed to measure the %R in the reduction state (corresponding to %Rmin), using a hydrazine solution at 35 wt% in water. In the case of electrochemical switching, ERDs were connected to a potentiostat and the reflectivity of the ERD active surface was recorded as a function of time and applied voltage. Optical switching times were determined for 95% of Δ%R. In the NIR, %R and Δ%R were measured at a single wavelength of 2.5 mm, while in the MIR %R and Δ%R were determined over a wavelength range from 3 to 5 mm in band II and from 8 to 12 mm in band III. Mechanical elongations were measured by a mechanical testing machine (UNIMAT plus 050-2 kN (ERICHSEN)) interfaced with a computer. Samples have typical dimension of Lwt ¼ 10 10 0.15 mm3. The samples were fixed between two clamps, drawn by a constant velocity (20 cm min 1) at room temperature, and the stress as a function of the strain was recorded. Young's modulus (E) was obtained from the slope in the linear part of the stress– strain curve.

Fig. 1. SEM pictures (cross-section) of NBR/POE/PEDOT (14 wt%) semi-IPN obtained after 1 h of polymerization.

3. Results and discussion

3.2. Optical characterization of the NBR/PEO/PEDOT semi-IPN

3.1. Elaboration and characterizations of the NBR/PEO/PEDOT semi-IPN

The maximum variations of reflectivity (Δ%R) in near infrared (NIR) and middle infrared (MIR) were evaluated. PEDOT on both faces of the film was oxidized and reduced by chemical process, as mentioned in Section 2. The reflectivity of NBR/PEO/PEDOT semiIPN in these two states was recorded in the 0.8–20 mm range by means of an integrating sphere coupled to the spectrophotometer. A semi-IPN containing 17 wt% of fully oxidized PEDOT exhibits reflective properties in the NIR and MIR ranges (Fig. 4) with a maximum reflectivity close to 60% beyond 8 mm. The drop in reflectivity below 2 mm corresponds to the plasma frequency usually observed in the NIR region for metallic materials [20,21] or conducting polymers [16,22,23]. In the neutral state, after chemical reduction of PEDOT, the reflectivity is low ( o20%). [22,24] Typically the results show the potentialities in term of reflectivity modulation ( 40% from 8 to 20 mm).

The NBR/PEO/PEDOT semi-IPNs were formed by EDOT polymerization within a pre-formed NBR/PEO IPN used as a flexible and soft host matrix for the ERD. The NBR/PEO IPN, whose synthesis and characterizations were reported previously [17], is a thin films (thickness¼140 mm) with a 54/46 (wt/wt) ratio of NBR/PEO partners. These IPNs have good mechanical properties due to the presence of NBR (elongation at break of 110% for a stress of 1.2 MPa) and high ionic conductivity promoted by the presence of the PEO network (10 3 S cm 1 at room temperature after swelling with 40% of EMITFSI). The PEDOT integration in the host matrix results from the dipping of EDOT swollen NBR/PEO IPNs (143 73 wt%) into a FeCl3 aqueous solution (1.5 mol L 1 at 40 1C) promoting the EDOT oxidative polymerization. The competition between EDOT desorption, FeCl3 absorption and EDOT polymerization within the host matrix leads to a PEDOT distribution only on a thin layer within both faces of the IPN film [16]. The specific PEDOT distribution in the semi-IPN has been evidenced by scanning electron microscopy (SEM) on the cross-section of a sample (Fig. 1) as already described [18,19]. Clearly, the semi-IPN behaves as a three-layer system as seen on the cross-section: two bright regions located on the outer part of the matrix containing the PEDOT interpenetrated into the NBR/PEO matrix, and a central part (darker region) containing the NBR and the PEO exclusively. The presence and localization of the PEDOT has been confirmed by SEM, coupled with EDAX measurements (Fig. 2). Indeed, Fig. 2A shows a homogeneous nitrogen atom distribution along the cross-section representative of the NBR distribution. On the other hand, sulfur atom distribution (Fig. 2B), exclusively labeling the PEDOT distribution, attests that PEDOT is present only in both semi-IPN film sides. The simultaneous presence of sulfur

and nitrogen atoms (Fig. 2C) at the edge of the cross-section indicates the PEDOT is well present within the hosting matrix. The NBR/PEO/PEDOT architecture is comparable to a three-layer structure (working-electrode|separator|counter-electrode) being a oneblock material. It is not possible to peel off the PEDOT layer from the film even by scraping, proving the efficiency of PEDOT interpenetration and the inability of delamination between the layers of the system over time. The PEDOT content within the IPN are directly controllable by the EDOT polymerization time. The PEDOT content increases from 3 to 17 wt% when the polymerization time increases from 10 to 75 min at 40 1C (Fig. 3).

3.3. Influence of the PEDOT content on the NIR contrast at 2.5 mm The reflectivity was monitored in the NIR at 2.5 mm versus the PEDOT content both in the chemically oxidized and reduced states of the semi-IPN (Fig. 5). In the reduced state, the amount of PEDOT has a weak influence on the reflectivity (%R2.5 mm red) and remains between 4% and 8%. On the contrary, the reflectivity of the oxidized semi-IPN (%R2.5 mm ox) is strongly dependent on the PEDOT content. It increases from about 10% to 30% when the PEDOT content increases from 3 to 7 wt%. Beyond 7 wt%, a plateau value at about %R2.5 mm ox ¼28% is reached. Therefore, the reflectivity variation Δ%R2.5 mm (Δ%R2.5 mm ¼%R2.5 mm ox %R2.5 mm red) increases with the PEDOT content and reaches a maximum close to 25% for a PEDOT content higher than 7 wt%. One can note that some samples could exhibit a reflectivity %R2.5 mm ox close to 35–38% which is the maximal reflectance measured at 2.5 mm by Heeger et al. [22] on free-standing oxidized PEDOT film.

36


Fig. 2. SEM-EDAX analyses of NBR/PEO/PEDOT (14 wt%) semi-IPN cross-section: nitrogen atom repartition (A), sulfur atom repartition (B), and distribution of nitrogen and sulfur atoms (C).

Fig. 3. PEDOT content ( ) and depth of PEDOT interpenetration ( ) versus polymerization time at 40 1C.

Fig. 4. Reflectance spectra from 0.8 to 20 μm of the active layer of a NBR/PEO/ PEDOT (17 wt%) semi-IPN oxidized by immersion in a NOBF4 solution ( ) and reduced by immersion in an hydrazine solution ( ).

Fig. 5. Evolution of the PEDOT reflectivity at 2.5 mm for semi-IPN structures in oxidized state ( ), in reduced state ( ) and reflectivity variation ( ) (Δ%R2.5 mm ¼% R2.5 mm_ox %R2.5 mm_red) versus PEDOT content.

Fig. 6. CV curves from 1st to 700th cycle, every 100th cycles of PEDOT interpenetrated in NBR/PEO IPN (7 wt% of PEDOT, scan rate of 50 mV s 1 in EMITFSI at room temperature).

3.4. Electrochemical characterization of the NBR/PEO/PEDOT semi-IPN The NBR/PEO/PEDOT samples were studied by cyclic voltammetry (CV) between 1.5 and þ1.5 V at a scan rate of 50 mV s 1 in a three-electrode electrochemical cell containing pure EMITFSI as electrolyte. The semi-IPN (0.83 cm 1.54 cm 131 mm, m¼15.1 mg) was pressed on a stainless steel grid and used as the workingelectrode. The CVs of a sample containing 7% of PEDOT are shown in Fig. 6 according to the number of cycles. During the first CV (dashed curve in Fig. 6), the PEDOT oxidation (p-doping, Scheme 1) occurs at 0 V since the current density is increasing to a peak value of

Scheme 1. Electrochemical oxidation/reduction process of PEDOT in NBR/PEO/ PEDOT systems.

7.1 mA cm 2 at 1.03 V. The process is followed by a diffusion plateau current typical of limitation by mass transport, i.e. counterions. During the backward scan, the reduction of p-doped PEDOT occurs with a cathodic peak current density jp,c ¼ 5.5 mA cm 2 at


Ep_c ¼ 0.62 V. The potential peak difference is quite high ΔEp ¼Ep, V and confirms that the electrochemical process is slow within IPN systems. A similar trend was already observed in NBR/PEDOT semi-IPN [19]. Depending on the mass of PEDOT in the sample (1.07 mg) and an estimated doping level δ¼0.2870.02 charge/monomer unit, the expected doping charge is Qd ¼2067 4 mC. From the integration of the first CV curve, we found Qd ¼0.187 C. The proportion of PEDOT which is electrochemically accessible inside the semi-IPN is 9076%. Cycling was performed in order to test the electrochemical stability until 15% of electroactivity loss. A 2% increase of PEDOT electro-activity is observed between the 1st and the 100th cycle. At the same time, and over the entire cycling test, the oxidation peak potential is shifting to lower potential values simultaneously with the increase of the reduction peak potential (Fig. 7A) resulting in a decrease of ΔEp. The trend suggests the charge transfer on PEDOT is becoming easier upon cycling. Looking at the doping charge Qd transferred during the repetitive cycles, PEDOT electroactivity decreases very slowly between the 100th cycle and the 700th cycle (Fig. 7B). Only 15% of redox (or doping) charge Qd is lost after 700 cycles, suggesting that the NBR/PEO/PEDOT system is quite stable.

aEp,c ¼1.65

3.5. Elaboration and characterizations of the electro-reflective devices After the synthesis of NBR/PEO/PEDOT semi-IPNs, the samples are swollen with EMITFSI used as an electrolyte to obtain the final

peak potential (V)

1.0

37

electroreflective device (ERD). The EMITFSI swelling degree in the semi-IPN is controlled by the immersion time. The reflectivity of the swollen semi-IPN can be monitored in the NIR at 2.5 mm depending on the swelling degree. Surprisingly, for a given sample, a systematic loss of reflectivity is observed regardless of the PEDOT content when the EMITFSI swelling degree increases (Fig. 8). One can suppose that the free volume of polymer chain increases arising from the swelling process, which leads to a spacing of the PEDOT chains. This phenomenon is due to the elastomeric nature of the hosting matrix and could result in a kind of dilution of the PEDOT on the semi-IPN surface. Therefore, the NIR optical properties are faded. The loss of optical properties is more pronounced with an increase of the PEDOT content in the semi-IPN, i.e. from 5 to 10 and 16 wt%. The reflectivity reaches a constant value from about 40 wt% EMITFSI swelling regardless of the PEDOT content. The control of oxidation states of the active surface was carried out by applying a low voltage ΔV, after clamping the swollen semiIPN between two metalized pieces connected to a potentiostat/ galvanostat. Both metalized pieces contain an open window as shown in Scheme 2, enabling the optical measurements of the obtained ERD. The configuration of the ERD is similar to an electrochemical cell, where the central NBR/PEO IPN swollen by EMITFSI ensures the ionic conductivity while the PEDOT on each side acts as optical electrodes. In such a device, the PEDOT is able to perform electrochemical and optical switches without the need for a current collector on the entire surface. The electrical connections were set in order to obtain the oxidation of the PEDOT on the active surface applying a positive voltage ΔV, while the PEDOT on the other side gets reduced and vice-versa when ΔV is negative. The ERDs were electrochemically switched between 71.5 V in order to obtain the maximum reflectivity levels in the NIR at a wavelength of 2.5 mm and in the MIR in bands II (3–5 mm) and III (8–12 mm).

0.5

0.0

-0.5 0

100

200

300

400

500

600

700

number of cycle

0.20

Fig. 8. Variation of NBR/PEO/PEDOT semi-IPN reflectivity as a function of EMITFSI swelling degree for samples with 5 wt% ( ), 10 wt% ( ) and 16 wt% ( ) of PEDOT content.

Charge density (C/cm²)

0.18 0.16 0.14 0.12 0.10 0.08 0.06 0.04 0.02 0.00 0

100

200

300

400

500

600

700

800

number of cycle Fig. 7. (A) Evolution of anodic and cathodic peak potential Ep,a Ep,c. (B) Evolution of the doping charge Qd as function of the number of cycles.

Scheme 2. Principle of fabrication of the electroreflective devices.

38


The variation of the reflectivity at a given wavelength λ (Δ% Rλ ¼ %R þ 1.5 V %R 1.5 V) is also affected by EMITFSI swelling degree. This can be evidenced by Fig. 9 and the trend is similar at higher wavelengths (MIR region). As first shown in Fig. 9, at low swelling degree (10 wt%) the electrochemical switching accompanied by an optical one is barely measureable (Δ%R2.5 mm ¼2%). This low modulation is related to the low ionic conductivity (3 10 5 S cm 1) for this swelling. Then, the Δ%R2.5 mm reaches a maximum value close to 18% when the swelling degree reaches between 20 and 30%. Finally, the Δ%R2.5 mm drops from 18 to 7% along with the EMITFSI content increases from 30 to 80 wt%. To possess the best Δ%R2.5 mm, the EMITFSI swelling degree of ERD should be set between 20 and 42 wt%. The reflectivity at 2.5 mm was also studied with different PEDOT loading for a given EMITFSI content (42 wt%). The reflectivity variation is also affected by the amount of PEDOT and gradually increases from Δ%R2.5 mm ¼8% to a maximum close to 15% for PEDOT contents higher than 8 wt% (Table 1). The result is explained by considering that the reflectivity will increase if the number of free carriers (responsible of the reflective behavior) close to the surface increases, in other terms, if the PEDOT content is high. However the switching time increases also with the PEDOT content (Table 1). For example, during oxidation, the optical switching time evolves from 42 s (6 wt% PEDOT) to 15 min (17 wt% of PEDOT). In the case of reduction, the trend is similar and switching time is increased from 1.2 to 64 min for the same devices samples. ERDs with PEDOT content of 8–12 wt% seem to be a good compromise between both acceptable reflectivity variation Δ%R and optical switching times. Regardless of the PEDOT content incorporated into the ERDs, the optical switching time is always shorter during oxidation than during reduction. In the case of our system, both surfaces (which contain the same amount of PEDOT) are identical since the device

is symmetrical but have opposite electrical state. Thus, when the active surface is oxidizing under applied potential, the opposite electrode (containing the same amount of electroactive PEDOT) is reduced at the same time and at the same rate. This is confirmed by the evolution of the absolute values of the charge density versus time, which are similar for the active surface oxidation and reduction (Fig. 10). Therefore, the charge/discharge kinetics does not explain the differences observed in optical switching time. Usually in the literature, the electrochemical reduction kinetics of doped ECP has been reported as slower than electrochemical oxidation kinetics [25–27]. Lacroix et al. [28] and other research groups [29–32] consider the ionic transport in the ECP is the limiting factor of the reduction. This explanation was moderated by Madden et al. [33] and Inganas et al. [34] who interpreted this phenomenon as a “moving front” depending on the metalinsulator transition. Therefore, it could be possible to explain why the optical switching is kinetically different from the electrochemical one. Looking at only one side of the device, and considering that the polymer is initially insulating (neutral state, optically low reflecting), the electrochemical charge/discharge switching yields information about the oxidized/reduced PEDOT amount into the whole PEDOT layer. In contrast, the reflection studies yield information on the optical state of the probed surface. The surface that is surface analyzed by the spectrophotometer corresponds to about 10% of the whole active surface. Scheme 3 shows the possible evolution of the metal-insulating transition of top surface layer for our ERD. Looking only at active surface (Scheme 3A) at the initial stage, PEDOT is insulating (dark blue) in NBR/PEO networks. The switching occurs only close to gold electrode contact where PEDOT is oxidized and get conducting following a nucleation process as described by Otero et al. [35]. Then, propagation of the moving

Fig. 9. Evolution of the reflectivity variation (Δ%R2.5 mm ¼%R þ 1.5 V %R 1.5 V) at 2.5 mm as a function of EMITFSI swelling degree for an ERD made with NBR/PEO/ PEDOT (17 wt%).

Table 1 Optical switching time for NBR/PEO/PEDOT semi-IPN swollen with 42 wt% ( 7 2 wt %) EMITFSI as a function of PEDOT content. PEDOT content Optical NIR reflectivity (wt%) variation Δ%R2.5 mm

6 8 12 15 17

7 15 12 14 16

Optical switching time (min) Oxidation ( þ 1.5 V)

Reduction ( 1.5 V)

0.7 1.7 2.5 6.6 15

1.2 2.9 4.5 10 64

Fig. 10. (A) Evolutions of the reflectivity ( ) and charge density ( ) versus time while the oxidation (þ 1.5 V) followed by the reduction ( 1.5 V) for a NBR/ PEO/PEDOT (17 wt%) swollen with 40 wt% EMITFSI and (B) superposition of the charge density absolute values versus time while the oxidation ( ) or the reduction ( ).


39

front occurs laterally as more and more PEDOT is conducting (Scheme 3B). Upon reverse switching, the moving front does not follow the same way only because the NBR/PEO/PEDOT volume is fully conducting (Scheme 3D, E and F). The ERD optical properties were extended in the MIR range (2.5–20 mm) for a sample with 17 wt% PEDOT and swollen with 45 wt% EMITFSI, which demonstrated good electro-reflective properties in the NIR. Fig. 11A shows the evolution of the optical reflectivity in this IR range when ΔV is scanned between þ1.5 and 2.5 V. When the applied voltage is þ 1.5 V, the active surface is reflective with an optical level up to 55% beyond 8 mm, which is very close to the one obtained with chemical doping (%R ¼60% see Fig. 4). Beyond þ 1.5 V, the reflectivity level remains unchanged. By decreasing the applied voltage by up to 1.5 V, the reflectivity is decreased from 55% to 20%. Nevertheless, a lower reflectivity close to 10% can be reached if the ΔV is further decreased to 2.5 V meaning that a more important de-doping state needs a highly negative voltage. The %R evolution was monitored in different spectral regions: bands II (3–5 mm) and III (8–12 mm) which are typical MIR regions of the thermal cameras. Fig. 11B shows the evolution of the reflectivity %R (2.5–20 mm) versus applied potential. Like %R varies from 6% to 26% at 2.5 mm, it changes from 6% to 32% in band II and from 10% to 51% in band III. Between þ 1.5 and 1.5 V, the 3 curves evolve similarly. Looking at only 2.5 mm, the ERD reaches its lower level of reflectivity for ΔV equal to or below 1.5 V. This is why experiments at 2.5 mm wavelength were performed with a ΔV range of 71.5 V. However, it is surprising to observe in bands II and III that the reflectivity level continues to decrease between 1.5 and 2.5 V. However beyond an applied potential of 1.5 V, the reversibility of the

Scheme 3. Representation of the insulating-to-metallic transition in the PEDOT from reduced to oxidized. (A) Full ERD cross section. (B–F) Active layer during commutation from reduced to oxidized PEDOT (A–C) and reverse commutation to reduced PEDOT.

Fig. 11. (A) Spectroelectrochemistry from 2.5 mm to 20 μm of a NBR/PEO/PEDOT semi-IPN (17 wt% of PEDOT) swelled by EMITFSI (45 wt%) for applied voltages from þ1.5 V to 2.5 V on the active layer and (B) integrated reflectivity at 2.5 mm ( ), in band II (3–5 μm) ( ) or band III (8–12 μm) ( ).

40


commutation drops, till 2.5 V where the devices did show no longer reversibility, as a proof of the degradation of the active layer. 3.6. Mechanical properties of the semi-IPN and ERDs ERDs based on IPN architecture reveal very interesting electrooptical properties, even quite similar to those reported in the literature for parent multilayered devices [14–16]. Moreover, ERDs based on NBR/PEOPEDOT semi-IPN possess very good mechanical properties. The ERDs are robust. The stress at break of ERDs swollen by 40 wt% of EMITFSI increases from 1.2 to 8.2 MPa with an elongation at break of between 84% and 14%, when the PEDOT amount increases from 0 to 19 wt% (Fig. 12). In comparison, PEO/ PEDOT ERDs containing 1 wt% PEDOT and swollen with 8 wt% EMITFSI showed a weaker mechanical strength with an elongation

Fig. 12. Stress versus strain for NBR/PEO/PEDOT semi-IPNs swollen by 40 wt% of EMITFSI and containing 0 wt% ( ), 6 wt% ( ), 12 wt% ( ) or 19 wt% ( ) of PEDOT.

at break below 10% for a stress at break lower than 1.4 MPa. These evolutions in the elongation and the stress at break result from the stiffness introduced by the PEDOT [36,37]. At the same time, the NBR/PEO/NBR ERD Young's modulus increases from 1.9 70.5 to 210 750 MPa with an increase of PEDOT content from 0 to 19 wt%. Furthermore, the devices are soft, flexible and do not break after bending and rolling up due to the mechanical properties given by the NBR network, making them a unique feature among the known ERDs (Fig. 13 and video attached).

4. Conclusion In this work, a PEDOT layer was interpenetrated in both faces of a soft and flexible thin film made of NBR/PEO IPN in order to obtain an optical electro-reflexive material in IR after addition of EMITFSI as an electrolyte. The device exhibits middle infrared and tunable optical properties between 2 and 20 mm upon electrical switching between 1.5 and þ1.5 V. The modulation of the reflectivity in bands II (3–5 mm) and III (8–12 mm) regions is 26% and 41% respectively. The incorporation of a soft synthetic rubber, the NBR, in the PEO/PEDOT semi-IPN material which was previously reported as the first ERD without the need for additional gold, ITO, or other extra layers because of the PEDOT acting both as a current collector and an active material, gives the possibility to design a flexible and robust free-standing ERD. The final device containing 40% EMITFSI shows elongations at break from 45% to 15% for stresses at break from 3 to 8 MPa depending on the amount of incorporated PEDOT.

Fig. 13. Pictures of NBR/PEO/PEDOT (15 wt%) swollen by 42% EMITFSI: full size before folding (A), firstly folded (B), secondly folded (C) and rolled up around a spatula (D).


41

Video S1. NBR/PEO/PEDOT-based IPN upon folding and mechanical stress.A video clip is available online. Supplementary material related to this article can be found online at http://dx.doi.org/10.1016/j.solmat.2014.03.036.

Acknowledgments This work has been supported by Délégation Générale de l'Armement. The authors acknowledge Serge Fanget from Lanxess Emulsion Rubber for graciously providing NBR, and Alexandre Khaldi from IEMN at the Université de Valenciennes et du Hainaut-Cambrésis (UVHC), France for SEM and EDAX measurements. Appendix A. Supplementary material The following are the supplementary data to this article: Video S1. References [1] T.A. Skotheim, R.L. Elsenbaumer, J.R. Reynolds, Handbook of Conducting Polymers, 3rd ed., New York, 2007. [2] P.M. Beaujuge, J.R. Reynolds, Color control in pi-conjugated organic polymers for use in electrochromic devices, Chem. Rev. 110 (2010) 268–320. [3] A. Bessiere, C. Marcel, M. Morcrette, J.M. Tarascon, V. Lucas, B. Viana, N. Baffier, Flexible electrochromic reflectance device based on tungsten oxide for infrared emissivity control, J. Appl. Phys. 91 (2002) 1589–1594. [4] C.G. Granqvist, Electrochromic devices, J. Eur. Ceram. Soc. 25 (2005) 2907–2912. [5] M.G. Hutchins, N.S. Butt, A.J. Topping, J. Gallego, P. Milne, D. Jeffrey, I. Brotherston, Infrared reflectance modulation in tungsten oxide based electrochromic devices, Electrochim. Acta 46 (2001) 1983–1988. [6] P.H. Aubert, A.A. Argun, A. Cirpan, D.B. Tanner, J.R. Reynolds, Microporous patterned electrodes for color-matched electrochromic polymer displays, Chem. Mater. 16 (2004) 2386–2393. [7] A.L. Dyer, C.R.G. Grenier, J.R. Reynolds, A poly(3,4-alkylenedioxythiophene) electrochromic variable optical attenuator with near-infrared reflectivity tuned independently of the visible region, Adv. Funct. Mater. 17 (2007) 1480–1486. [8] P. Chandrasekhar, T.J. Dooley, Far-IR transparency and dynamic infrared signature control with novel conducting polymer systems, Proc. SPIE 2528 (1995) 2528. [9] H. Pagès, P. Topart, D. Lemordant, Wide band electrochromic displays based on thin conducting polymer films, Electrochim. Acta 46 (2001) 2137–2143. [10] I. Schwendeman, J. Hwang, D.M. Welsh, D.B. Tanner, J.R. Reynolds, Combined visible and infrared electrochromism using dual polymer devices, Adv. Mater. 13 (2001) 634–637. [11] P. Chandrasekhar, B.J. Zay, G.C. Birur, S. Rawal, E.A. Pierson, L. Kauder, T. Swanson, Large, switchable electrochromism in the visible through farinfrared in conducting polymer devices, Adv. Funct. Mater. 12 (2002) 95–103.

[12] A.A. Argun, P.H. Aubert, B.C. Thompson, I. Schwendeman, C.L. Gaupp, J. Hwang, N.J. Pinto, D.B. Tanner, A.G. MacDiarmid, J.R. Reynolds, Multicolored electrochromism polymers: structures and devices, Chem. Mater. 16 (2004) 4401–4412. [13] L.H. Sperling, D. Klemper, L.A. Utracki, L.H. Sperling, D. Klemper, L.A. Utracki (Eds.), Interpenetrating Polymer Networks, American Chemical Society, Washington, 1991. [14] F. Tran-Van, L. Beouch, F. Vidal, P. Yammine, D. Teyssie, C. Chevrot, Selfsupported semi-interpenetrating polymer networks for new design of electrochromic devices, Electrochim. Acta 53 (2008) 4336–4343. [15] M. Mallouki, P.-H. Aubert, L. Beouch, F. Vidal, C. Chevrot, Symmetrical electrochromic device from poly(3,4-(2,2-dimethylpropylenedioxy)thiophene)-based semi-interpenetrating polymer network, Synth. Met. 162 (2012) 1903–1911. [16] P. Verge, P.H. Aubert, F. Vidal, L. Sauques, F. Tran-Van, S. Peralta, D. Teyssie, C. Chevrot, New prospects in the conception of IR electro-tunable devices: the use of conducting semi-interpenetrating polymer network architecture, Chem. Mater. 22 (2010) 4539–4547. [17] L.J. Goujon, A. Khaldi, A. Maziz, C. Plesse, G.T.M. Nguyen, P.H. Aubert, F. Vidal, C. Chevrot, D. Teyssie, Flexible solid polymer electrolytes based on nitrile butadiene rubber/poly(ethylene oxide) interpenetrating polymer networks containing either LiTFSI or EMITFSI, Macromolecules 44 (2011) 9683–9691. [18] C. Plesse, A. Khaldi, Q. Wang, E. Cattan, D. Teyssié, C. Chevrot, F. Vidal, Polyethylene oxide–polytetrahydrofurane–PEDOT conducting interpenetrating polymer networks for high speed actuators, Smart Mater. Struct. 20 (2011) 124002. [19] I. Fabre-Francke, P.-H. Aubert, S. Alfonsi, F. Vidal, L. Sauques, C. Chevrot, Electropolymerization of 3,4-ethylenedioxythiophene within an insulating nitrile butadiene rubber network: application to electroreflective surfaces and devices, Sol. Energy Mater. Sol. Cells 99 (2012) 109–115. [20] S.H. Brewer, S. Franzen, Indium tin oxide plasma frequency dependence on sheet resistance and surface adlayers determined by reflectance FTIR spectroscopy, J. Phys. Chem. B 106 (2002) 12986–12992. [21] G.B. Smith, A. Maaroof, A. Dowd, A. Gentle, M. Cortie, Tuning plasma frequency for improved solar control glazing using mesoporous nanostructures, in: A. Gombert (Ed.), Photonics Eur., International Society for Optics and Photonics, 2006, p. 61970T–61970T–9. [22] Y.H. Chang, K. Lee, R. Kiebooms, A. Aleshin, A.J. Heeger, Reflectance of conducting poly(3,4-ethylenedioxythiophene), Synth. Met. 105 (1999) 203–206. [23] J. Joo, E.J. Oh, G. Min, A.G. Macdiarmid, A.J. Epstein, Evolution of the conducting state of polyaniline from localized to mesoscopic metallic to intrinsic metallic regimes, Synth. Met. 69 (1995) 251–254. [24] A.N. Aleshin, R. Kiebooms, A.J. Heeger, Metallic conductivity of highly doped poly(3,4-ethylenedioxythiophene), Synth. Met. 101 (1999) 369–370. [25] Y. Tezuka, K. Aoki, K. Shinozaki, Synth. Met. 30 (1989) 369–379. [26] Y. Tezuka, T. Kimura, T. Ishii, K. Aoki, J. Electroanal. Chem. 385 (1995) 51–55. [27] Y. Tezuka, K. Aoki, H. Yajima, T. Ishii, J. Electroanal. Chem. 425 (1997) 167–172. [28] J.C. Lacroix, K. Fraoua, P.C. Lacaze, Moving front phenomena in the switching of conductive polymers, J. Electroanal. Chem. 444 (1998) 83–93. [29] F. Miomandre, M.N. Bussac, E. Vieil, L. Zuppiroli, Influence of the local electric field on ionic transport during redox switching of conducting polymers, Chem. Phys. 255 (2000) 291–300.

42


[30] X.Z. Wang, B. Shapiro, E. Smela, Development of a model for charge transport in conjugated polymers, J. Phys. Chem. C 113 (2009) 382–401. [31] X.Z. Wang, B. Shapiro, E. Smela, Visualizing ion currents in conjugated polymers, Adv. Mater. 16 (2004) 1605. [32] X.Z. Wang, E. Smela, B. Shapiro, Understanding ion transport in conjugated polymers, in: Y. BarCohen (Ed.), Smart Structures and Materials 2004: Electroactive Polymer Actuators and Devices, 2004, pp. 146–154. [33] M.R. Warren, J.D. Madden, Electrochemical switching of conducting polymers: a variable resistance transmission line model, J. Electroanal. Chem. 590 (2006) 76–81.

[34] T. Johansson, N.K. Persson, O. Inganas, Moving redox fronts in conjugated polymers studies from lateral electrochemistry in polythiophenes, J. Electrochem. Soc. 151 (2004) E119–E124. [35] T.F. Otero, I. Boyano, Nucleation and nonstoichiometry in electrochromic conducting polymers, Chemphyschem 4 (2003) 868–872. [36] C. Plesse, F. Vidal, H. Randriamahazaka, D. Teyssie, C. Chevrot, Synthesis and characterization of conducting interpenetrating polymer networks for new actuators, Polymer 46 (2005) 7771–7778. [37] N. Festin, A. Maziz, C. Plesse, D. Teyssie, C. Chevrot, F. Vidal, Robust solid polymer electrolyte for conducting IPN actuators, Smart Mater. Struct. 22 (2013).