Sustainable, Flexible, and Biocompatible Enhanced Piezoelectric Chitosan Thin Film for Compliant Piezosensors for Human Health

The necessity to continuously and seamlessly monitor human health is calling for compliant, comfortable, and safe wearables. The employment of piezoelectric biopolymers in form of thin film perfectly matches with these needs due to their inherent flexibility, sensitivity, and biocompatibility. Among them, chitosan is a low cost, highly sustainable, and biocompatible material with a great potential for applications in compliant wearables. However, chitosan and biopolymers in general show relatively low piezoelectric coefficients and processing difficulties. Here, it is shown a facile approach to increase more than twice the piezoelectric coefficient of thin chitosan film and to process this promising biomaterial for the fabrication of the first set of thin chitosan film‐based ultrasound transducers. This work leverages the exploitation of environmental‐friendly biopolymers in the development of compliant wearable transducers and thus represents a step forward in the development of completely biodegradable, transparent, thin, and flexible piezoelectric macro‐ and microtransducers.


Introduction
The growing need for systems to closely and continuously monitor human health requires the development of compliant applications. [8] Nevertheless, until now they have been poorly exploited for the development of thin film piezoelectric devices because of their general low piezoelectric coefficient and difficulties in manufacturing, since biopolymers are not compatible with common microfabrication techniques and chemicals. [9] Several fabrication attempts employing biopolymers have been performed for the development of Micro Electro Mechanical Systems (MEMS) and ultrasound transducers (UT), but they display poor performances linked to the material processing. Miniaturized UT have been developed with thin silk films showing good piezoelectric properties, but the device results to be rigid and opaque. [10] Moreover, in the context of developing implantable devices, silk has shown issues regarding immunogenicity and rare allergic reactions. [11] Other devices have been developed with Poly Lactic-co-Glycolic Acid and Poly L-Lactic Acid (PLLA), but complex procedures had to be developed for poling and for processing the material in nanofibers for obtaining good performances. [12,13] In general, biopolymers are difficult to process to fabricate flexible MEMS and miniaturized devices [9] , because they cannot be easily produced in form of thin films, and they are not always compatible with the tools and chemicals used in microfabrication technology. Both flexibility and miniaturization are necessary features in piezoelectric MEMS to cover a large range of working frequencies and applications that span from physiological monitoring to ultrasound sensing. [14][15][16] To circumvent all these limitations, we focused our attention on chitosan, a low-cost natural biopolymer derived from chitin, the second most abundant polysaccharide on earth. It represents a highly sustainable resource that can be extracted from the exoskeleton of crustaceans from food industrial waste. Chitosan is a nonimmunogenic carbohydrate, [17] largely used for biomedical applications, such as cellular scaffold or drug delivery platforms, by virtue of its high biocompatibility and antibacterial properties. [18][19][20] It can be processed in a very thin and highly compliant flexible film (Figure 1a). Moreover, thanks to its noncentrosymmetric crystal structure (Figure 1b,c), chitosan exhibits piezoelectric behavior when processed with simple solvent-casting technique without the need of the expensive poling procedure, making this material very intriguing to be exploited for batch production. [21] However, like biopolymer in general, chitosan shows a relatively low piezoelectric coefficient around 6 pC N -1, [22] which can hinder its applications for health monitoring devices. Second, at the best of our knowledge, the available literature lacks protocols to process thin chitosan films to be employed for the development of MEMS and compliant wearable devices. The only report of flexible chitosan based piezoelectric patch employs a glycine-chitosan composite, where chitosan acts just as a nonpiezoelectric matrix, while glycine crystals require the orientation of their piezoelectric domains. [23] In this work, we demonstrate the enhancement of the piezoelectric properties of chitosan thin films by an inexpensive and simple chemical treatment and we propose tailored processing protocols to enable the development of the first compliant piezoelectric chitosan-based miniaturized transducers. Thanks to the tuned crystallographic alignment obtained by controlled simple solvent casting technique and post-casting conditions, we obtained a transparent and 15 µm thick chitosan film, whose Piezoresponse Force Microscopy (PFM) measurements revealed a piezoelectric coefficient of 15 pC N -1 -more than doubling the current reference value of 6 pC N -1 . [22] This advancement, together with tailored processing protocols, has enabled us to develop three different classes of flexible, compliant and sustainable transducers with additional characteristics required for wearable devices: i) a highly sensitive chitosan flag-like transducer with silver/gold metal electrodes (hereinafter called CT-Au/Ag); ii) a transparent chitosan-based flag-like transducer with poly(3,4-ethylenedioxythiophene) polystyrene sulfonate (PEDOT:PSS) -Glycerol based electrodes (PG) (hereinafter called CT-PG); iii) a miniaturized and semitransparent circular piezoelectric Micromachined Ultrasound Transducer (pMUT) based on chitosan and gold/PG electrode.

Fabrication and Characterization of Chitosan Film with Enhanced Piezoelectricity
Chitosan films are fabricated according to a simple and green solvent casting method [21] with controlled casting and postcasting conditions. A water solution with 1% (w/v) Chitosan and 1%(v/v) Lactic Acid is stirred, poured in a petri dish and left overnight at 40 °C under a vacuum of 200 mbar. After solvent evaporation, films are immersed in a bath with a NaOH 1 m solution for neutralization. It is reported that few seconds variations of neutralization time enhance crystallinity of chitosan film. [24] Starting from this assumption, we hypothesized that the highest crystallinity of the polymeric mesh can improve piezoelectric performances of the film. In order to study the effect of longer neutralization times on the polymer chains Figure 1. a) Chitosan is processed in order to obtain thin and transparent films with enhanced piezoelectric behavior. b) Chitosan derives from the N-deacetylation of Chitin, a renewable raw material that can be extracted from the exoskeleton of crustaceans, and it is composed by repetitions of two units of N-Acetyl-D-Glucosamine and D-Glucosamine. c) Simplified scheme of the supramolecular organization of chitosan chains arranged in a noncentrosymmetric cell with possible positions of water molecules that stabilize the structure by column-like junctions. Red atoms represent Oxygen, green atoms represent Carbon and blue atoms represent Nitrogen. Hydrogen atoms in chitosan have been omitted.
www.advelectronicmat.de organization, four different films were obtained starting from the same amount of chitosan solution and then were soaked in the basic solution for 10, 30, 60, and 90 min (samples will be indicated hereafter as 10F, 30F, 60F, and 90F, respectively). After neutralization, films are rinsed with distilled water, peeled off from petri dishes and placed on the substrate to dry for Piezoresponse Force Microscopy (PFM) measurements. The substrate used is a glass with a solution based on PEDOT:PSS spin coated and cured on it: this ensures a flat and conductive bottom electrode. PFM measurements were performed on all the samples in order to evaluate the piezoelectric properties of the films. From Figure 2 it is possible to compare the amplitude of the displacement (Figure 2a-d) and the phase (i.e., the piezoelectric domain orientation) (Figure 2e-h) for all the samples, highlighting an increase of displacement that follows the increase in neutralization time. By PFM spectroscopy the samples are locally excited at different AC voltages (from 0.5 to 5 V) and the corresponding displacement due to DC sweep (from −10 to +10 V) on each point is measured, obtaining the curves whose slopes represent an estimation of the effective piezoelectric coefficient d 33 eff (Figure 2i-l). Values obtained were 7.40 pC N -1 for 10F, 14.36 pC N -1 for 30F, 15.56 pC N -1 for 60F, and 9.13 pC N -1 for 90F indicating that the neutralization treatment enhances the piezoelectric behavior of the chitosan films. In particular, 60 min of neutralization represent the best time duration choice in terms of piezoelectric response www.advelectronicmat.de as piezoelectric coefficient d 33 eff is increased more than twice, compared to literature. [22] Additionally, the thickness of films decreases with the neutralization time ( Table 1), in good agreement with literature. [25] Atomic force microscopy (AFM) measurement revealed a very flat surface with a medium roughness of 3 nm (RMS) for samples 10F, 30F, and 60F and a slight increase of 4.45 nm for sample 90F ( Figure S1a-d, Supporting Information). These roughness values match well with those in literature for flexible electronics. [26] Observation of samples with polarized optical microscopy (POM) (Figure 2m-p) gives information on the spherulite structures, which represent the nucleation center for film formation during solvent casting. It is possible to observe them in all the samples, but only in the case of sample 60F there is a consistent presence of Maltese crosses in the center of the spherulites, indicating a high order of polymer chains [27] that can be related to the enhanced piezoelectric effect.
The structural characterization of the films after NaOH treatment was carried out using the X-ray diffraction (XRD) technique. Diffractograms were acquired both in reflection and in transmission geometry. The resulting curves are displayed in Figure 3a,b. The patterns obtained in reflection geometry (Figure 3a) show the presence of two broad peaks at around 2θ = 10° and 20°, characteristic for the (020) and (110) crystallographic planes, respectively, of the hydrated form of chitosan with orthorhombic cell (space group: P2 1 2 1 2 1 ), which represents the most common polymorphism of chitosan. [28,29] The complete absence of the (020) peak, in all the acquired patterns in transmission geometry (Figure 3b), reveals the presence of a strong preferred orientation related to the (020) planes, which are therefore parallel to the film surface. In this measurement geometry, the incident X-ray beam is orthogonal to the film surface and the (020) planes are never in Bragg condition with the beam. In contrast, the crystallographic planes tilted with respect to the film surface give rise to the (110) diffraction peak and, visible at around 2θ = 21° as a weak shoulder, the (120) peak. The presence of such a pronounced preferred orientation makes unreliable the quantification of the crystallinity, typically calculated using the integrated intensity of crystalline peaks (obtained by pattern refinement) and the integrated intensity of the halo representing the amorphous fraction of the sample. [30] Instead, an interesting parameter is represented by the ratio between the integrated intensities of the (020)  www.advelectronicmat.de peak and the amorphous halo. Table 1 shows how this value is higher for sample 60F. This parameter is not influenced by the sample's thickness and can be associated with crystallinity or with a preferential orientation: a higher value (Table 1) means a higher crystalline order. These findings, when compared to the piezoelectric coefficient measurements, underline how the marked intensity from plane (020) can be related to a higher piezoelectric performance. It was previously observed in literature that an orthorhombic crystal cell is linked to improved piezoelectric performances in ceramic piezoelectric materials. [31] In fact, in the case of chitosan films, the order of the plane (020), parallel to the film surface, after 60 min of NaOH solution treatment may result in a better charge separation. Thermo-gravimetric analysis (TGA) performed on the samples revealed that the NaOH treatment reduces the degradation temperature (Figure 3c,d). This is probably due to the loss of water molecules in the crystal cell that not only makes the film more compact, but also causes a rearrangement in the crystal organization creating an unstable structure. [25,32] This instability can explain the highest piezoelectric behavior of the sample 60F that represents an equilibrium point in which domains are free to rotate during the displacement of the material, molecular planes are more compact inducing better charge separation and the crystal cell shows a strong orthorhombic order that promotes piezoelectricity. [31] After 60 min of neutralization, the film becomes too compact, the domains cannot rotate freely and the crystal cell loses its shape. This explanation is in good agreement with dynamic mechanical analysis (DMA) too (Table 1). In fact, the Young's modulus of 90F is higher, revealing a more brittle and fragile film compared to the other samples ( Figure S2, Supporting Information).

Fabrication and Characterization of High-Sensitivity Chitosan Flag-Like UT (CT-Au/Ag)
In order to explore the fabrication possibilities and test the properties of the highly piezoelectric thin films, we realized chitosan miniaturized flags embedded between metal electrodes with an active area of 8 mm × 8 mm. The fabrication process (Figure 4a) starts with the printing of the bottom electrode made of a silver-based nanoparticles-ink (AgCite) deposited on glass by means of DragonFly LDM System 3D printer. Then, a 200 nm thick layer of PEDOT:PSS blended with Glycerol (4% w/w) (PG) is spin-coated on the bottom electrode to promote the adhesion with chitosan. The blend PG is chosen because glycerol in this concentration enhances conductivity of PEDOT:PSS. [33] The bottom electrode is cured and then chitosan film is deposited onto it obtaining samples for each chitosan film (10F, 30F, 60F, 90F). Once dried, in order to shape the top electrode, soft adhesive Kapton shadow masks are cut with Laser Cutter VLS2.30DT and attached on the chitosan film after a proper alignment with the bottom electrode. Then, top gold electrode is e-beam evaporated directly on the chitosan film through the shadow masks. By virtue of the strong adhesion between chitosan film and bottom electrode, the device is easily peeled off from the glass substrate in order to have freestanding flags. Finally, the device is wired and encapsulated with a thin conformal Parylene-C layer (Figure 4b,c). The device was connected directly to the oscilloscope and two stimuli with different intensities, i.e., finger tapping and blowing, were exerted on the device in order to qualitatively assess its ability to register physiological low frequency events. Plots with amplitude versus time are reported ( Figure S3a,b, Supporting Information) showing the signal recorder by CT-Au/ Ag. The characterization of the piezoelectric output under a bending stimulus is an important aspect to assess the flexibility of the device. [34] Therefore, a three-point bending test was performed on the device while connected to the oscilloscope and the piezoelectric signal under bending stimulus was recorded ( Figure S4; Supporting Information).
An electrical characterization was performed measuring the capacitance of the final devices in a frequency range from 2 Hz to 1 MHz. All samples show capacitance in the range of nF and in all cases the capacitance decreases with frequency ( Figure 4d; Figure S6a-c, Supporting Information). Dielectric permittivity ε r is estimated with the formula ε r = ε 1 /ε 0 = Cd/ε 0 A, where ε 0 is the vacuum permittivity (8.85 × 10 −12 F m -1 ), d is the thickness of the chitosan film, A is the active area's surface and C is the measured capacitance. Values obtained are shown in Table 2.
As for standard capacitors, the capacitance decreases with frequency and the low dielectric permittivity at 1 MHz indicates that the film displays high performances at ultrasonic range demonstrating the high potential as UT. [23] Moreover, comparing the dielectric permittivity with piezoelectric coefficient, it is shown that the sample 60F shows the lowest dielectric permittivity value and the highest d 33 when compared to other samples ( Figure 4e). The combination of low permittivity and high d 33 is typical of materials that show a high efficiency in electro-mechanical conversion. A figure of merit (FOM) such as FOM = (d 33 ) 2 /ε 0 ε r can therefore give a clear indication on how good is the piezoelectric material to transduce between mechanical and electrical energy. This FOM is equal to 0.67 GPa -1 for 10F, 2.3 GPa -1 for 30F, 5.6 GPa -1 for 60F and 1.74 GPa -1 for 90F ( Figure 4f, Table 2). Thus, among the four treatment applied, chitosan film treated for 60 min with NaOH shows the highest FOM, also indicating that its performances are comparable to or even better than other synthetic piezoelectric materials. [35] A calibration of the devices was performed by an impact test using small plummets with different weights that can free fall from 15 cm height. In Figure 4g-j the outcome of the calibration for each sample is showed: all samples show a linear response range between 0.04 and 0.16 kPa, which is a standard working range of ultrasound transducer. [16] Only in case of devices based on chitosan 30F and 60F, the response arises between 0 and 0.5 kPa. The sensitivity is calculated as the slope of the linear fit and the sample 60F shows the highest value of 80 mV kPa -1 (Figure 4i, Figure S6b, Supporting Information), together with the maximum output that reaches 32 mV ( Figure S6a, Supporting Information). At the best of our knowledge, the sensitivity value of the sensor is in the highest range when compared to previously reported piezoelectric flexible transducers, both biodegradable and nonbiodegradable ( Table 3). These results confirm the better performances of the 60F film as expected from the material characterization, compared to films produced with different times of treatment. For this reason, all the devices described hereinafter are fabricated and tested using film 60F. According to the calculated sensitivity of www.advelectronicmat.de 80 mV kPa -1 , the amplitude of the finger tapping and blowing signals recorded with CT-Au/Ag ( Figure S3a,b, Supporting Information) are obtained by the application of a pressure respectively of 1.25 kPa and 25 Pa, which are included into the physiological range described in literature for moderate finger tapping [36,37] and human breathing or blowing. [38,39]

Fabrication and Characterization of Transparent and Organic Chitosan Flag-like UT (CT-PG)
After complete characterization of CT-Au/Ag, a second version of the flag-like devices was developed using only the blend of PG as both bottom and top electrodes. The fabrication approach (Figure 5a) starts by depositing a thin layer of Parylene-C on a  www.advelectronicmat.de glass substrate as the poor adhesion of the Parylene-C to glass allows the detachment of the entire device from glass after the last fabrication steps. Then, a layer of PG is spin coated on Parylene-C and cured as bottom electrode and a chitosan film 60F is deposited on the bottom electrode and let dry. Top electrode is made by spin coating and curing PG on a polypropylene substrate. PEDOT:PSS thin layers can be moved using water as transfer medium, [45] but, in our case, a dry method was used because of the high liquid absorbance of chitosan film. Using polyimide tape, it is possible to remove the top electrode from the polypropylene substrate and transfer it on the chitosan film, exploiting the poor adhesion of PG layer to polypropylene. Finally, the device is peeled off from glass substrate and conformally covered with Parylene-C (Figure 5b,c). The device was connected directly to the oscilloscope and two stimuli with different intensities, i.e., finger tapping and blowing, were exerted on the device in order to qualitatively assess its ability to register physiological low frequency events with nonmetallic electrodes. Plots with amplitude versus time are reported ( Figure S7a are consistent with the physiological pressure ranges reported in literature for these types of stimuli. [36][37][38][39] A 3-point bending test was performed on the device while connected to the oscilloscope and the piezoelectric signal under bending stimulus was recorded ( Figure S8, Supporting Information). These devices are characterized by transparency and total biodegradability as both chitosan and PEDOT:PSS show these properties. [46] In order to assess the transparency of CT-PG device, transmittance analysis was performed by spectrophotometer and the result, for all layers of the device, is shown in Figure 5d. The entire device shows an average transmittance of 50% in the visible range (between 400 and 800 nm).

Fabrication and Characterization of Flexible and Semi-Transparent pMUT Based on Biodegradable Chitosan
In order to fabricate the pMUT and achieve suspended chitosanbased membranes (Figure 6a), a Poly(dimethylsiloxane) (PDMS) substrate was used where cavities underneath the suspended membranes for pMUTs were obtained by means of molding technique. [47] The mold is fabricated with SU-8 on a silicon wafer by standard UV lithography. It consists of a row of cylindrical pillars with 25 µm height and 100 µm base diameter. Then, a mixture of curing agent and PDMS (ratio 1:10) is poured on the mold, spin coated and cured in order to obtain its positive version. The resulting PDMS structure consists of a row of cylindrical cavities with 5 µm-thick PDMS structural membrane on top (as in the schematic of Figure 6a). Bottom electrode is fabricated by spin coating and curing of PG layer (200 nm thick) on PDMS and then chitosan film 60F is deposited onto it and left to dry. In order to shape the top electrode, soft adhesive Kapton shadow masks are cut by laser with apertures of 100 µm diameter and attached on the chitosan film after a proper alignment with the circular openings of the cavities. As a final step, a gold layer of 100 nm is e-beam evaporated directly on the chitosan film through the shadow mask (final result is displayed in Figure 6a,b).
In order to mechanically characterize the fabricated device in terms of out-of-plane main flexural mode resonances, pMUT

www.advelectronicmat.de
resonator was placed on a piezoelectric commercial disk, which was actuated with a voltage sweep from 2 MHz until 3 MHz at +10 V, exploiting the reverse piezoelectric effect. In this way, the induced mechanical vibrations in the disk can be transferred to the pMUT resonator that vibrates accordingly, showing its own flexural resonances at proper matching frequencies. The resonances were achieved by measuring the out-of-plane displacement of the pMUT membrane by a Laser Doppler Vibrometer (LDV). The corresponding displacement versus frequency plots reveals a more intense peak of vibration at 2.25 MHz (Figure 6c). Moreover, the out-of-plane deflection shows that the membrane, which is assumed to behave as a fully clamped device, has a Gaussian-like shape (where the maximum displacement is achieved at the center of the membrane) that corresponds to the first fundamental vibration mode (0,1) of the membrane (Figure 6d). [48,49]

Chitosan Devices Testing in Ultrasonic Environment
The ability of the device in sensing ultrasound waves was tested with an ad hoc setup (Figure 7a). It consists of a pulser/receiver (DPR300 Pulser-Receiver from JSR Ultrasonics) connected to an immersion commercial transducer (bandwidth 1 MHz, nominal central frequency 1 MHz, NTDXducer V303-SU, Olympus) that provides to the commercial transducer a negative driving pulse with a controlled amplitude and a pulse repetition frequency (PRF) of 5 kHz (the pulse duration is 30-40 ns). A commercial hydrophone (flat response up to 25 MHz) was used as reference. The incoming acoustic waves were collected by an oscilloscope, able to collect both reference and chitosan-based devices response as a voltage output from the pulser/receiver. In particular, the chitosan-based devices signal was pre-amplified by 36 dB at a frequency of 1 MHz by a custom wideband amplifier with high input impedance of 1 MΩ before being sent as input to the pulser/receiver, with input impedance of 500 Ω and gain of 30 dB. The commercial UT emitter, the reference hydrophone and the chitosan-based devices were immersed in a tank filled with water as medium for the ultrasound propagation. All the tested devices were able to detect ultrasound incoming pulse waves centered at 1 MHz with an amplitude V p-p = 5.3 mV for CT-Au/ag, V p-p = 3.8 mV for CT-PG and V p-p = 5.4 mV for pMUT after amplification (Figure 7b-d). To preliminary study the maximum sensing distance, multiple recordings have been performed while moving the devices back and forth from the UT emitter by using an X-Y translational stage. The point target focus of the transducer lays in a range between 1.5 cm and 2 cm, though CT-Au/Ag, CT-PG and pMUT can detect ultrasound up to 10 cm far from the emitter that corresponds to a Time-Of-Flight of 70 µs (Figure 7e-g), calculated considering the speed of sound in water of 1428 m/s.

Biocompatibility
A study on the biocompatibility of the materials used to develop the chitosan-based ultrasound transducers was performed in order to assess the potentiality of these devices to be applied as wearable or implantable sensors. A preliminary viability test was performed on each layers of the devices in order to demonstrate that all the constituent materials are biocompatible and safe and, as well as each step of the fabrication protocol does not lead hazards for human health. In accordance to ISO 10993-5 (2009), [50] which contains guidelines for biological evaluation of medical devices, we employed murine 3T3 embryonic fibroblasts, cultured in 96 wells plates with Dulbecco's modified

www.advelectronicmat.de
Eagle's medium (DMEM) enriched with Fetal Bovin Serum, Penicillin/Streptomycin and Glutamine. After 24 h, small pieces of 2 mm × 2 mm for each layer configuration from all the three devices ( Table 4) have been inserted into each well, in order to put the materials in contact with the cellular medium. Cell viability was assessed by absorbance at 450 nm at 24 and 72 h after the first contact with cellular medium through a Cell Counting Kit-8 that measures the cellular metabolic activity by conversion of a water-soluble tetrazolium salt into a formazan dye. Results are shown in Figure 8. At 24 h we observed nonsignificant difference if compared to negative control for most samples excluding number 6 and 7, in which cellular viability increases. At 72 h we observed a general decrease in cellular viability, which indicates that molecules released from the samples interacted with cells. Nonetheless, the percentage of viable cells both at 24 and 72 h is higher than 70% which, in compliance to ISO 10993-5:2009, means that all tested materials show no cellular toxicity. Moreover, since it is known that silver-based inks can be harmful to cells, [51,52] we suggest a positive influence of chitosan film in retaining silver nanoparticles on its surface in samples 3 and 4, leading to a minimum release of silver, that, in turn, results in a nontoxic cellular effect. In fact, in sample 5, which contains silver nanoink confined by Parylene-C, we observed a nonsignificant reduction of viability. The strong interaction between chitosan and silver was already observed in literature [53] and it is explained by the high affinity of primary amino groups in chitosan with positively charged metal ions.

Conclusions
The importance of the development of compliant transducers based on flexible, green and biocompatible materials and sustainable fabrication processes is essential both for human health and environment safety. The present work introduces a combined simple strategy to both improve the piezoelectric properties of chitosan and design innovative fabrication approaches in order to exploit it for a novel class of efficient, flexible and compliant wearable piezoelectric devices and skin electronics. Our results demonstrate that 60 min neutralization of chitosan film with NaOH represents a simple and green strategy to increase the piezoelectric performances of the In addition, we provide tailored fabrication protocols to process this biodegradable material in order to develop the first prototypes of flexible UTs with a high sensitivity of 80 mV kPa -1 , transparent UTs and flexible and semi-transparent pMUT based on chitosan and working in a range of few MHz. Aging tests and further experiments will be performed to assess the piezoelectric behavior of the devices under extreme environ-mental conditions such as low and high temperature. We therefore envision our work as a step forward toward innovative green compliant transducers.

Experimental Section
Preparation of Chitosan Piezoelectric Film: Chitosan (1% w/v, M.W. 890000 avg.) (Glentham) was mixed with Lactic Acid (1% v/v) (Sigma-Aldrich) in a water solution (100 mL). The solution was stirred until the polymer was completely dissolved. The solution was centrifuged at 3500 rpm for 10 min to remove impurities. Then, part of the solution (20 mL) was poured into glass petri dish 10 cm Ø and let dry overnight at 40 °C and 200 mbar. After solvent evaporation, films were immersed in an acqueous solution of NaOH 1 m for different amount of time (10′, 30′, 60′ and 90′) in order to study the influence of this step on the piezoelectric behavior of the film. After required time was passed, NaOH solution was discarded and the films were washed with distilled water until the pH of the washing solution was around 7 checking by litmus paper (VWR). After washing, the films were peeled off from the petri dish, placed on a curved glass surface, rolled on the substrate and let dry overnight under the hood. All the steps were performed in a clean room in order to avoid powder contamination.
Piezoresponse Force Microscopy (PFM): Samples were prepared starting from the substrates. Glass for printing with a thickness of 300 µm (Nanodimension) was released from the plastic cover, cut into pieces of 2 cm × 2 cm and treated with oxygen plasma (RFG300 Deiner, Ebhausen, Germany) at 120 W for 10′. Plasma oxygen ensures a good bond with the conductive ink. For the bottom electrode, a blend of ink PEDOT:PSS high conductivity grade (Sigma-Aldrich) and Glycerol (4% w/w) (Sigma-Aldrich) was stirred overnight as Glycerol enhances the conductivity of PEDOT:PSS, [54] improve flexibility and adhesion to chitosan film. The mixture was then spin coated on the glass in order to obtain a layer 200 nm thick and cured at 140 °C for 5′. Then, a second layer of PEDOT:PSS and Glycerol was spin coated on the first one and cured at 140 °C for 30′. Wet chitosan film, after washing, were placed on the PEDOT:PSS/Glycerol substrate leaving part of the electrode exposed and let dry overnight. Then, the sample was cut into pieces of 1 cm × 1 cm, attached to a magnetic plate and contacted with silver paste (RS). The samples as prepared can be placed on the sample holder of Piezoresponse Force Microscopy (CSI Nano-Observer AFM microscope) for measurements. The tip employed for measurements was a conductive ANSCM-PC tip (AppNano) with Pt coating, length 450 µm and spring constant 0.02-0.8 N/m used in contact mode with a frequency around 100 kHz. An alternating current at 5 V was applied to the sample during scanning in order to obtain Amplitude and Phase images. The estimation of piezoelectric coefficient was performed by tip calibration with a Lithium Niobate calibration sample.
Atomic Force Microscopy analysis: AFM measurements were performed with the same instrument for PFM by using a silicon tip FORT (AppNano) with length 225 µm and spring constant 0.6-3.7 N/m. Polarized Optical Microscopy: POM was exerted with polarization filters on a Nikon Eclipse L200NSD Microscope.
X-Ray Diffraction analysis: XRD patterns of chitosan films were recorded at room temperature on a Rigaku SmartLab X-Ray powder diffractometer equipped with a 9 kW CuKα rotating anode, operating at 40 kV and 150 mA. A Göbel mirror was used to convert the divergent X-ray beam into a parallel beam and to suppress the Cu Kβ radiation. The measurements in reflection mode were carried out using a Parallel Slit Analyzer (PSA) with angular aperture = 0.5° and 0D scintillation detector. The patterns in transmission mode were acquired using a D\teX Ultra 1D silicon strip detector. For calculation of (020)/amorphous ratio, line profile analysis (LPA), using HighScore plus 5.0 software from PANalytical, was carried out to obtain the integrated intensity from the (020) peak and the amorphous halo. The split pseudo-Voigt function was used as peak profile function while the background was modeled with a polynomial function.  Mechanical Test: Traction tests were performed with Dynamic Mechanical Analyzer (DMA Q800, TA Instruments, USA) on chitosan specimens sized 1 × 6 cm with a constant speed 1 N min −1 until failure.
CT-Au/Ag Fabrication: The bottom electrode was a 1 µm thick layer of silver-based nanoink AgCite printed on 300 µm thick glass (Nano Dimension, Israel) in the shape of a "P" by means of Nano Dimension's DragonFly LDM System 3D printer (Israel). The mixture of PEDOT:PSS and Glycerol previously described was spin coated on the electrode and cured at 140 °C for 30′. Then chitosan film was deposited on it and let dry overnight. A thin layer of Gold 100 nm was e-beam evaporated after 30 nm of chromium for adhesion thanks to adhesive kapton mask shaped with laser cutter. The device can be peeled off the glass, connections were created with copper tape and wires and, finally, the entire device was conformally covered with a 1 µm thick Parylene-C layer at room temperature.
3-Point Bending Test: The bending tests were performed with Dynamic Mechanical Analyzer (DMA Q800, TA Instruments, USA) with the dedicated 3-point bending tool on the devices connected to the oscilloscope. The tool applies a repetitive pressure of 0.02 N with frequency of 2 Hz and displacement of 500 µm.
Electrical Characterization: Capacitance measurements were performed connecting the device to an LCR-meter Agilent E4980A. In order to scan the capacitance in a range of frequencies, LCR meter was connected to a pc and driven by a MatLab script.
Impact Test: For calibration of device sensitivity a homemade impact test setup was built. A trampoline 15 cm high was built and the device was positioned perpendicularly under the trampoline edge. Then, a metallic plate with known size was attached to a nylon wire and positioned on the top electrode of the device, while the other end of the wire was fixed to the edge of the trampoline. Plummets with known weight were left to free fall from the edge of the trampoline on the top of the device using the wire as a guidance in order to hit the device in the center. The device was connected to an oscilloscope (Tektronix MDO 4104-3) to register electrical signal due to the impacts of the plummets on the device.
CT-PG Fabrication: Microscope glasses (Thermo Scientific) were covered with a 1 µm thick Parylene layer, then the blend of PEDOT:PSS and Glycerol was spin coated on it and cured at 140 °C for 30′ as bottom electrode. Chitosan film was deposited onto it and let dry overnight. Top electrode was realized by spin coating a double layer of PEDOT:PSS/ Glycerol mix on a polypropylene substrate with Kapton mask for the shape and curing at 140 °C for 30′. After curing, the top electrode was transferred on adhesive Kapton tape and then transferred from the tape directly on the chitosan film overnight dried. The device can be detached from the glass thanks to the Parylene substrate that can be eliminated from the device by peeling, connections were made with copper tape and wires and, lastly, a conformal layer of Parylene-C 1 µm thick was deposited on the device.
Spectrophotometer Analysis: Transmittance analysis were performed with a Cary 5000 UV-vis-NIR Spectrophotometer in a range from 400 to 800 nm on incremental layers of the device: Parylene coating only; Parylene coating and bottom PEDOT:PSS/Glycerol electrode; Parylene, bottom electrode and chitosan; Parylene, bottom, chitosan and top electrode. Every sample was averaged from 3 to 5 times.
pMUT fabrication: The mold was made by SU-8 2025 (MicroChem, Westborough, MA, USA) a negative photoresist. Briefly, SU-8 was spin coated on silicon wafer at 3000 rpm for 30″ and it was soft baked at 65 °C/3′ and 95 °C/5′. Then, using a soft maks (JD Photo Data), cylindrical pillars were made by photolithography with Mask Aligner SUSS MA8/BA8 (Garching, Germany) with exposure 165 mJ/cm 2 . After exposure, the sample was post exposure baked at 65 °C/1′ and 95 °C/5′ and developed for 1′ in SU-8 developer. Then sample development was then stopped in IPA and dried under a nitrogen flow. In the end, the sample was hard baked at 95 °C for 1 h. PDMS (Sylgard 184, Dowsil), in ration 1:10 with curing agent, was poured on the SU-8 mold and spin coated at 3800 rpm for 60″ in order to obtain a 25 µm high structure and 5 µm thick rounded membranes on top of SU-8 pillars. Then, PDMS was cured at 90 °C for 10′. PDMS was treated with oxygen plasma (RFG300 Deiner, Ebhausen, Germany) at 150W/5′ and then a double layer of the mix of PEDOT:PSS/Glycerol was spin coated and cured at 140 °C/30′ before depositing chitosan film and let to dry. Adhesive polymide masks were cut by laser cutter (VLS2.30DT, Universal Laser Systems), to obtain the shape of the top electrode and were attached to the chitosan surface by using a custom-made mask aligner. Gold top electrode 100 nm thick was e-beam evaporated after a 30 nm thick chromium layer on the chitosan and then Kapton mask were removed. Electrical connections were made with copper tape and wires and the device was conformally coated with Parylene-C.
pMUT Central Frequency Characterization: The device was positioned on top of a commercial piezoelectric disk (P.I., USA) that was connected to a Laser Doppler Vibrometer (LDV Polytec MSA-500 Micro System Analyzer, Polytec, Germany) and actuated with a current sweep until 4 MHz in order to give a mechanical excitation to the membrane of the device and measures its displacement thanks to LDV reading system.
Ultrasound Testing Setup: DPR300 Pulser/Receiver (JSR, Japan) and Ultrasound transducer 1 MHz center frequency (NtdXducer, USA) was used as ultrasound origin with pulser at 475 V and 90.2 µJ per pulse. A hydrophone (flat response up to 25 MHz, Onda, USA) was used as reference. Chitosan device was amplified by a custom made amplifier and both device and hydrophone were connected to the oscilloscope (Tektronix MDO 4104-3) to read ultrasounds. Measurements were performed in a tank filled with distilled water with devices at different distances from the ultrasound source.
Cells were seeded in 96 well microplates at a density of 5000 cells/ well and incubated for 24 h for cell adhesion; subsequent incubation made with 2×2 mm samples UVC sterilized and inserted into each well. After 24 and 72 h, a 10 µL aliquot of WST-8 solution was added to each well. The 96-well microplates were incubated for 3 h in a humidified atmosphere of 5% CO 2 and 37 °C. Subsequently, the absorbance of orange WST-8 formazan product was read through Tecan M200 Infinite Microplate Reader (Tecan, Switzerland) at 450 nm. Measurements were performed in 8 replicates and statistic T-student test was applied between each sample and negative control with significance level at 0.05.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.