*2.1. Materials*

The preforms for permeability measurements were obtained using balanced and unidirectional carbon fiber fabrics and unidirectional (UD) tapes. The preforms investigated in this work are shown in Figure 1 and their properties are reported in Table 1.

(Rheometric Scientific).

*Materials* **2020**, *13*, x FOR PEER REVIEW 4 of 17

**Figure 1.** Preforms for permeability measurements: (**a**,**d**) balanced, (**b**,**e**) unidirectional (UD) stitched and (**c**,**f**) unidirectional (UD) consolidated by AFP. **Figure 1.** Preforms for permeability measurements: (**a**,**d**) balanced, (**b**,**e**) unidirectional (UD) stitched and (**c**,**f**) unidirectional (UD) consolidated by AFP.


**Table 1.** Carbon fiber preform configuration. **Table 1.** Carbon fiber preform configuration.

(mm<sup>3</sup> ) 80×50 × 8 80 × 50 × 8 80 × 50 × 8 Preform manufacturing process vacuum bagging vacuum bagging automated fiber placement Preform A was a balanced fabric, produced by Hexcel (Stanford, CA, USA) with the trade name G0926 HS06K. It was a 0.38-mm-thick 5H satin with HEXTOW AS4C GP 6K with a nominal weight of 375 g/m<sup>2</sup> , the same weight distribution in warp and weft directions and filled with an epoxy binder. The binder content was 4% of the weight of the carbon fiber [15]. Preforms A were obtained in laboratory by vacuum bagging of [0]22 stacks of G0926 5H satin fabric in an oven during the following Preform A was a balanced fabric, produced by Hexcel (Stanford, CA, USA) with the trade name G0926 HS06K. It was a 0.38-mm-thick 5H satin with HEXTOW AS4C GP 6K with a nominal weight of 375 g/m<sup>2</sup> , the same weight distribution in warp and weft directions and filled with an epoxy binder. The binder content was 4% of the weight of the carbon fiber [15]. Preforms A were obtained in laboratory by vacuum bagging of [0]<sup>22</sup> stacks of G0926 5H satin fabric in an oven during the following thermal cycle: heating up to 100 ◦C at 3 ◦C/min, isotherm for 75 min at 110 ◦C and cooling to room temperature at 1.5 ◦C/min. This cycle was able to dissolve the epoxy powders used as binder and to give stiffness to the preforms.

thermal cycle: heating up to 100 °C at 3 °C/min, isotherm for 75 min at 110 °C and cooling to room temperature at 1.5 °C/min. This cycle was able to dissolve the epoxy powders used as binder and to give stiffness to the preforms. Preform B was a stitched unidirectional carbon fabric, produced by Solvay S.A. (Bruxelles, Preform B was a stitched unidirectional carbon fabric, produced by Solvay S.A. (Bruxelles, Belgium) with the trade name BNFC-24k IMS-(0)-196-600, obtained with IM65 carbon fibers containing a binder for preform fabrication. Preforms A and B, were provided by Leonardo SpA (Foggia, Italy), according to the same consolidation cycle.

Belgium) with the trade name BNFC-24k IMS-(0)-196-600, obtained with IM65 carbon fibers containing a binder for preform fabrication. Preforms A and B, were provided by Leonardo SpA (Foggia, Italy), according to the same consolidation cycle. Preform C was made of an unidirectional TX 1100 IMS65 24 k fabric, produced by Solvay S.A. Preform C was made of an unidirectional TX 1100 IMS65 24 k fabric, produced by Solvay S.A. (Bruxelles, Belgium) with IM65 carbon fibers. Preforms C, provided by Novotech Aerospace Advanced Technology srl (Avetrana, Italy), were prepared by automated fiber placement (AFP) using a tape width of 8 mm.

(Bruxelles, Belgium) with IM65 carbon fibers. Preforms C, provided by Novotech Aerospace Advanced Technology srl (Avetrana, Italy), were prepared by automated fiber placement (AFP) using a tape width of 8 mm. Permeability experiments were performed at room temperature using polyethylene glycol 400 (PEG400) as test fluid provided by Sigma-Aldrich (Milano, Italy). This latter has been chosen since its viscosity value at room temperature is in the viscosity range of a thermosetting resin used in resin Permeability experiments were performed at room temperature using polyethylene glycol 400 (PEG400) as test fluid provided by Sigma-Aldrich (Milano, Italy). This latter has been chosen since its viscosity value at room temperature is in the viscosity range of a thermosetting resin used in resin infusion processes at high temperature. The viscosity of the test fluid was characterized by rheological analysis carried out in an ARES parallel plate rheometer with a 50-mm plate diameter (Rheometric Scientific).

infusion processes at high temperature. The viscosity of the test fluid was characterized by

#### *2.2. Experimental Set-Up for Permeability Measurements by Ultrasonic Wave Propagation Materials* **2020**, *13*, x FOR PEER REVIEW 5 of 17

The experimental set-up for out-of-plane permeability measurements by ultrasonic wave propagation is sketched not in scale in Figure 2. The carbon fiber (CF) preform was placed between two glass plates, each of them containing a hole for the inlet and outlet of the impregnation fluid. The thickness of the glass plates was 8 mm and 20 mm for the lower and upper plate, respectively. The higher thickness was chosen in order to avoid overlapping echoes due to the reflections from the glass plate and the wet carbon fiber preform. In order to provide the vacuum tightness of the measurement system, a silicone seal was used. A resin distribution net at the top and bottom side of the sample allowed a through-thickness unidirectional flow under a constant injection pressure, obtained by a Vacuum Assisted Resin Infusion (VARI). Fiber volume fraction was changed by adjusting the distance between the glass plates. After turning on the vacuum pump, the inlet valve was opened leading the fluid to flow from the tank through the preform. Since the measurement was carried out in unsteady conditions, an unsaturated permeability was determined. During the ultrasonic measurement, the balance—framed by a red dotted line in Figure 2—was not used. *2.2. Experimental Set-Up for Permeability Measurements by Ultrasonic Wave Propagation*  The experimental set-up for out-of-plane permeability measurements by ultrasonic wave propagation is sketched not in scale in Figure 2. The carbon fiber (CF) preform was placed between two glass plates, each of them containing a hole for the inlet and outlet of the impregnation fluid. The thickness of the glass plates was 8 mm and 20 mm for the lower and upper plate, respectively. The higher thickness was chosen in order to avoid overlapping echoes due to the reflections from the glass plate and the wet carbon fiber preform. In order to provide the vacuum tightness of the measurement system, a silicone seal was used. A resin distribution net at the top and bottom side of the sample allowed a through-thickness unidirectional flow under a constant injection pressure, obtained by a Vacuum Assisted Resin Infusion (VARI). Fiber volume fraction was changed by adjusting the distance between the glass plates. After turning on the vacuum pump, the inlet valve was opened leading the fluid to flow from the tank through the preform. Since the measurement was carried out in unsteady conditions, an unsaturated permeability was determined. During the ultrasonic measurement, the balance—framed by a red dotted line in Figure 2—was not used.

**Figure 2.** Experimental set-up for out-of-plane permeability measurements by ultrasonic wave propagation: (**a**) general sketch not in scale; (**b**) particular of the ultrasonic measurement device. **Figure 2.** Experimental set-up for out-of-plane permeability measurements by ultrasonic wave propagation: (**a**) general sketch not in scale; (**b**) particular of the ultrasonic measurement device.

An ultrasonic transducer (A109-R Olympus Italia srl, Milano, Italy, center frequency =5 MHz, active diameter =13 mm) was fixed on the upper glass plate by a probe housing system capable to keep a constant force on the transducer. The ultrasonic frequency has been chosen as the best compromise among the wave attenuation and resolution, which increase with frequency, and the echo damping, which decreases with frequency and can lead to the overlapping of echoes related to different interfaces. An ultrasonic couplant gel was used to provide the transmission of sound energy between the transducer and the glass plate. The ultrasonic transducer, connected with a pulser-receiver (Sofratest, Ecquevilly, France), worked in pulse-echo mode, acting as emitter and receiver of ultrasonic waves at the same time [55]. The transducer was excited by a broadband pulser with an electric voltage of 200 V and a pulse duration of 200 nanoseconds. The echoes, sampled at a frequency of 60 MHz, were composed of 2048 points. The received signals, after digitalization, were automatically recorded each 0.15 sand displayed by a custom made software. Ultrasonic waves were continuously sent through the preform and it was possible to measure their time-of-flight that is the time taken for the wave to pass the preform and come back to the transducer after reflection at interfaces characterized by different acoustic impedances [56]. The two first round-trip echoes in the wet preform were considered. The ultrasonic measurement device has been set-up for carbon fiber preforms up to 10 mm. However, thicker preforms can be investigated by increasing the thickness of the upper glass plate.
