Non-Invasive Approaches for the Evaluation of the Functionalization of Melamine Foams with In-Situ Synthesized Silver Nanoparticles

The use of polymeric nanocomposites has arisen as a promising solution to take advantage of the properties of nanoparticles (NPs) in diverse applications (e.g., water treatment, catalysis), while overcoming the drawbacks of free-standing nanoparticles (e.g., aggregation or accidental release). In most of the cases, the amount and size of the NPs will affect the stability of the composite as well as their performance. Therefore, a detailed characterization of the NPs present on the nanocomposites, including their quantification, is of vital importance for the optimization of these systems. However, the determination of the NPs load is often carried out by destructive techniques such as TGA or ICP-OES, the development of non-invasive approaches to that aim being necessary. In this work, the amount of silver NPs synthesized directly on the surface of melamine (ME) foams is studied using two non-invasive approaches: colorimetry and X-ray radiography. The obtained results show that the amount of silver NPs can be successfully determined from the luminosity and global color changes of the surface of the foams, as well as from the X-ray attenuance.


Introduction
In the past decades, the use of porous polymer structures as host materials for the production of nanocomposites has become a relevant field of research [1][2][3][4][5][6][7][8]. Among the most effective approaches to fabricate porous polymer nanocomposites, the following can be mentioned: mechanical blending [2,7,9] and surface functionalization processes [8,10] such as dip-coating [11,12], spray-coating [5], or in situ synthesis [3,6,10]. These procedures aim to take advantage of the features of both matrix and filler of the nanocomposites. On the one hand, the porous polymer provides a large surface area and excellent mechanical strength for long-term use. On the other hand, the anchoring of the nanoparticles (NPs) to the matrix avoids risks of their accidental release to the environment [4,13,14], as well as their efficiency loss in some applications (e.g., water treatment, catalysis) due to aggregation [6,15,16].
However, a critical factor in keeping the functionality of the filler is a proper selection of the nanocomposite fabrication route. First, the choice of the porous polymeric matrix is strongly linked to the application. The wide availability of diverse polymers allows providing matrices with optimal features for each application, in terms of the porous structure, mechanical and thermal behavior, wettability, chemical stability, or even bio-compatibility or bio-degradability [1,4,17,18]. Second, the Polymers 2020, 12, 996 3 of 15 use of nanofillers (nanoclays and nanosilicas) produce a significant enhancement in the nucleation mechanism, reducing cell size and increasing cell density. Besides, Prade et al. [37] developed a method to study the orientation of the fibers in short fiber-reinforced polymer composites based on X-ray vector radiography, showing the potential of this technique for the non-invasive characterization of this kind of nanocomposites. In addition, Roels and Carmeliet [38] employed X-ray radiography to measure the process of moisture transport in cellular materials. They obtained accurate data on the moisture content and proved the applicability of the method on two different materials with different structures, homogeneous and heterogeneous.
Herein, facile non-invasive approaches based on colorimetry and X-ray radiography to determine the amount of noble metal NPs synthesized in porous polymer foams are presented. First, AgNPs were in situ synthesized in commercial porous melamine (ME) foams following the approach proposed by Pinto et al. [10]. A wide set of ME foams and AgNPs nanocomposites (ME/Ag) was obtained using two AgNO 3 solutions (0.85 and 1.70 mg/mL) and different reaction times (from 1 to 7 days). Then, the obtained ME/Ag foams were analyzed by different approaches to determine the amount of AgNPs synthesized on the foams. In particular, two methodologies based on colorimetry and one on X-ray radiography were successfully tested. All these non-invasive approaches were calibrated using reference values previously reported by using a destructive ICP-based method [10], and provided accurate results on the entire set of ME/Ag foams.

Materials
Commercial melamine (ME) foams with high porosity (0.994), open pores, and mean pore size below 500 µm were kindly provided by Flexicel Industrial S.L. (Spain). Silver nitrate (AgNO 3 ) and tetrahydrofuran (THF) were purchased from Scharlab S.L. (Spain). Deionized water was employed for the preparation of the silver nanoparticles (AgNPs) precursor solutions as well as for the rinsing of the foams.

Functionalization of the ME Foams
The in situ synthesis of silver nanoparticles (AgNPs) on the surface of pure melamine (ME) foams (samples 2 × 1 × 1 cm 3 ) was achieved following the approach proposed by Pinto et al. [10] which provides a rather constant AgNPs distribution along the foams. The pure foams were immersed in 20 mL solutions with two different AgNO 3 concentrations (0.85 and 1.70 mg/mL in a H 2 O/THF mixture (1:19 v:v)) and shook for different times ranging from 1 to 7 days. The AgNPs synthesis is expected to happen following a redox reaction involving the oxidation of -NH 2 groups of the melamine and the reduction of the Ag + ions ( 3Ag . After the reaction time was reached, the samples were washed for five cycles, each cycle involved immersing the foam in 20 mL of distilled water, and shaking it for 2 min. Finally, all the samples were dried at room temperature.

Experimental Techniques
The structure and morphology of the ME foams, before and after the synthesis of the AgNPs were studied by scanning electron microscopy (SEM) (HITACHI FlexSEM 1000) after being coated with gold (10 nm) and by optical microscopy using a Microscopy EduBlue at 40×. SEM micrographs were obtained using an accelerating voltage of 10 kV and a working distance of 8 mm.
The determination of the amount of AgNPs synthesized (wt.%) into the foams was attempted by different methodologies, using the values previously reported by Pinto et al. [10] (measured by inductive couple plasma-optical emission spectroscopy (ICP-OES)) as reference values (see Section S1 in the Supplementary Information for more details). It should be noticed that only values for reaction times of 1, 3, 5, and 7 days were previously reported. Thermo gravimetric analysis (TGA) (SDTA851, Mettler Toledo) was performed with a heating ramp from 50 to 800 • C at a pace of 20 • C/min with nitrogen atmosphere (60 mL/min), followed by an isotherm at 800 • C for 15 min with air atmosphere (200 mL/min). Three samples of about 2 mg were extracted from different regions of each foam and analyzed by TGA, whereas the amount of AgNPs on the ME/Ag foams was estimated from the residue obtained for the ME and the ME/Ag foams (see Section S1 in the Supplementary Information for more details). Color measurement was carried out using a commercial color reader Colourpin SE (NCS Color AB, USA) using a standard illumination D65, 10 • standard observer, and achieving a precision ∆E* ab < 0.1. This color reader provides the color coordinates of the sample on the standard color space CIE L*a*b*, where there are tree axes orthogonal to each other: measurable lightness L* (values range from 0 (black) to 100 (white)); and two chromatic axes a* (values from green (negative) to red (positive)), and b* (goes from blue (negative) to yellow (positive)). The mean values and standard deviation of each coordinate were obtained from at least three measurements of each sample carried out both in external surfaces of the samples and internal surfaces exposed by cutting. The color differences between the pure ME foams and the obtained ME/Ag foams were evaluated in terms of the change of the luminosity (∆L*, Equation (1)) and CIE L*a*b* color change (∆E* ab , Equation (2)) which takes into account the three color coordinates.
X-ray radiographs were obtained using a constructed X-ray equipment consisting of a low energy micro-focus X-ray source (L10101, Hamamatsu. Voltage: 20-100 kV, Current: 0-200 µA), and a high sensitivity flat panel detector (C7940DK-02, Hamamatsu. 2240 × 2344 pixels, 50 µm pixel size) [9,36,39]. Sections with thickness about 2.8-2.9 mm were taken from the foams to perform these measurements. The radiographs were acquired with a tube voltage and current fixed at (40 kV, 170 µA) and an exposure time of 1500 ms. In order to enhance the contrast, each image was the result of integrating four consecutive images. The linear attenuation coefficient, µ (m −1 ), of each foam has been calculated after applying logarithmic conversion to Beer-Lambert's attenuation law (Equation (3)). This equation relates to the intensity transmitted through the material, I, with the initial X-ray beam intensity, I 0 , and the sample's thickness, t.
It is known that the attenuation of a two-phase system is a linear combination of that of every present phase. In particular, the attenuation coefficient of a material (µ) is a function of the number concentration, C (mol/m 3 ), of the present species and their X-ray cross-section, σ (m 2 ) (Equation (4)). Since the ME/Ag samples were obtained from the same initial foam without significant alteration of the polymer matrix, the value of µ ME will be the same for the treated foams and the reference pure foam. Meanwhile, the second term in Equation (4) (µ Ag ) will change depending on the amount of AgNPs present on the ME/Ag foams.
Thus, it should be possible to obtain the attenuation coefficient because of the sole presence of different amounts of AgNPs (µ Ag ) for every foam from the measured µ and µ ME values [40].
Finally, the comparison between the different techniques employed was carried out by calculating the mean squared error (MSE) of the values provides for each method with respect to the trends obtained from the ICP-OES values employed as a reference (Equation (5)). where n is the number of samples, Y i is the value obtained from the different methods andŶ i is the value previously provided by the ICP-OES measurements [10].

Characterization of the ME and ME/Ag Foams
It is well-known that the amount of NPs transferred to the foam following this in situ synthesis approach depends mainly on the reaction time and the concentration of precursors. Therefore, in order to have an appropriate set of samples for this study, ME/Ag foams were prepared using two different AgNO 3 concentrations (0.85 and 1.70 mg/mL) and seven reaction times for each concentration (1 to 7 days). These samples are expected to provide a wide range of AgNPs loads in which the proposed techniques to determine the amount of NPs (wt.%) can be validated. In the first step, the porous structure of the produced foams was analyzed by SEM microscopy ( Figure 1).
As previously reported by Pinto et al. [10], no significant differences were observed in the porous structure of the ME foams after being immersed in the AgNO 3 solutions with different concentrations (0.85 and 1.70 mg/mL) for several days (from 1 to 7). In fact, porous structure of the functionalized foams presents no changes even after 7 days of reaction time, independently of the concentration of the precursor solution ( Figure 1, more details in the Supplementary Information, Section S2). Moreover, focusing on the ME surfaces, it was possible to identify in the functionalized foams the presence of aggregates of AgNPs with sizes around 200-300 nm ( Figure 1, Figures S2-S4). However, previous studies proved that, in addition to these aggregates, these surfaces are homogeneously covered by individual AgNPs with average sizes below 10 nm [10].
where n is the number of samples, Yi is the value obtained from the different methods and Ŷi is the value previously provided by the ICP-OES measurements [10].

Characterization of the ME and ME/Ag Foams
It is well-known that the amount of NPs transferred to the foam following this in situ synthesis approach depends mainly on the reaction time and the concentration of precursors. Therefore, in order to have an appropriate set of samples for this study, ME/Ag foams were prepared using two different AgNO3 concentrations (0.85 and 1.70 mg/mL) and seven reaction times for each concentration (1 to 7 days). These samples are expected to provide a wide range of AgNPs loads in which the proposed techniques to determine the amount of NPs (wt.%) can be validated. In the first step, the porous structure of the produced foams was analyzed by SEM microscopy ( Figure 1).
As previously reported by Pinto et al. [10], no significant differences were observed in the porous structure of the ME foams after being immersed in the AgNO3 solutions with different concentrations (0.85 and 1.70 mg/mL) for several days (from 1 to 7). In fact, porous structure of the functionalized foams presents no changes even after 7 days of reaction time, independently of the concentration of the precursor solution ( Figure 1, more details in the Supplementary Information, Section S2). Moreover, focusing on the ME surfaces, it was possible to identify in the functionalized foams the presence of aggregates of AgNPs with sizes around 200-300 nm (Figures 1, S2-S4). However, previous studies proved that, in addition to these aggregates, these surfaces are homogeneously covered by individual AgNPs with average sizes below 10 nm [10].   The influence of the preparation parameters of ME/Ag foams and the AgNPs distribution was also studied by optical microscopy (Figures 2 and 3). Regardless of the precursor concentration employed, the functionalization of ME foams with the AgNO 3 solutions induced an evident change in the color of the samples, starting after the first day. The color changes from light gray for the pure ME foams to light brown on the ME/Ag foams obtained after one day. This brown color is a clear indication of the presence of AgNPs [10].
In addition, optical micrographs of the entire set of samples (Figures 3 and 4) show that the foams become darker as the reaction time increases. These micrographs, showing a large area of the samples, prove that the functionalized foams present a quite homogeneous color, which suggests that a homogeneous AgNPs coating was obtained. Furthermore, it is known that the amount of NPs increases with the reaction time; therefore, the observed progressive change in color is a qualitative indication of the relationship between both magnitudes. The influence of the preparation parameters of ME/Ag foams and the AgNPs distribution was also studied by optical microscopy (Figures 2 and 3). Regardless of the precursor concentration employed, the functionalization of ME foams with the AgNO3 solutions induced an evident change in the color of the samples, starting after the first day. The color changes from light gray for the pure ME foams to light brown on the ME/Ag foams obtained after one day. This brown color is a clear indication of the presence of AgNPs [10].
In addition, optical micrographs of the entire set of samples (Figures 3 and 4) show that the foams become darker as the reaction time increases. These micrographs, showing a large area of the samples, prove that the functionalized foams present a quite homogeneous color, which suggests that a homogeneous AgNPs coating was obtained. Furthermore, it is known that the amount of NPs increases with the reaction time; therefore, the observed progressive change in color is a qualitative indication of the relationship between both magnitudes.   The influence of the preparation parameters of ME/Ag foams and the AgNPs distribution was also studied by optical microscopy (Figures 2 and 3). Regardless of the precursor concentration employed, the functionalization of ME foams with the AgNO3 solutions induced an evident change in the color of the samples, starting after the first day. The color changes from light gray for the pure ME foams to light brown on the ME/Ag foams obtained after one day. This brown color is a clear indication of the presence of AgNPs [10].
In addition, optical micrographs of the entire set of samples (Figures 3 and 4) show that the foams become darker as the reaction time increases. These micrographs, showing a large area of the samples, prove that the functionalized foams present a quite homogeneous color, which suggests that a homogeneous AgNPs coating was obtained. Furthermore, it is known that the amount of NPs increases with the reaction time; therefore, the observed progressive change in color is a qualitative indication of the relationship between both magnitudes.     [10] in the functionalized ME/Ag foams obtained after different reaction times in two concentrations of AgNO3.

TGA Study of ME/Ag Foams
First, the amount of Au transferred to ME/Ag foams was attempted to be determined by TGA analysis, a conventional and destructive method for this kind of characterization (see Section S1 in the Supplementary Information for more details). Figure 4 shows the obtained results compared with the results previously published using ICP-OES analysis [10]. A wide range of contents (1 to 18 wt.%) is obtained indicating that the preparation process was successful in producing materials with enough differences in composition. In general, the samples with lower content of NPs (less than 5 wt.%) do not follow the expected trends. At the same time, at medium and high contents, the data adjust better to the trends, although with significant deviations regarding the values obtained in several cases. These discrepancies can have two origins. On the one hand, the TGA technique has an experimental error of 1% (according to the technical specifications of the measurement system), which implies errors of up to 33% in the estimation of the AgNPs contents at low values (TGA residues about 11 wt.%), while only about 6% at the higher contents (TGA residues about 25 wt.%). On the other hand, the determination of the AgNPs amount from the TGA residue was performed assuming a constant ME residue after the TGA program (about 8 wt.%), being not possible to discard that in the presence of AgNPs the ME degradation behaves differently.

Colorimetry Study of ME/Ag Foams
As shown in Figures 2 and 3, the immersion of the ME foams into AgNO3 solutions induced a color change on the surface of the foams, which seemed to be related to the amount of AgNPs present on the foams. Thus, a colorimetric analysis was performed to determine accurately the color coordinates in each foam. Then, this information can be related to the preparation conditions (i.e., reaction time and precursor solution concentration) to check if it allows monitoring the amount of AgNPs synthesized on the foams. Average values and standard deviations of the colorimetric measurements are shown in Table 1. Table 1. Colorimetry CIE L*a*b* coordinates obtained for the pure ME foam and ME/Ag foams produced using 0.85 and 1.70 mg/mL precursor solutions as well as different reaction times from 1 to 7 days.

Samples
Reaction time

TGA Study of ME/Ag Foams
First, the amount of Au transferred to ME/Ag foams was attempted to be determined by TGA analysis, a conventional and destructive method for this kind of characterization (see Section S1 in the Supplementary Information for more details). Figure 4 shows the obtained results compared with the results previously published using ICP-OES analysis [10]. A wide range of contents (1 to 18 wt.%) is obtained indicating that the preparation process was successful in producing materials with enough differences in composition. In general, the samples with lower content of NPs (less than 5 wt.%) do not follow the expected trends. At the same time, at medium and high contents, the data adjust better to the trends, although with significant deviations regarding the values obtained in several cases. These discrepancies can have two origins. On the one hand, the TGA technique has an experimental error of 1% (according to the technical specifications of the measurement system), which implies errors of up to 33% in the estimation of the AgNPs contents at low values (TGA residues about 11 wt.%), while only about 6% at the higher contents (TGA residues about 25 wt.%). On the other hand, the determination of the AgNPs amount from the TGA residue was performed assuming a constant ME residue after the TGA program (about 8 wt.%), being not possible to discard that in the presence of AgNPs the ME degradation behaves differently.

Colorimetry Study of ME/Ag Foams
As shown in Figures 2 and 3, the immersion of the ME foams into AgNO 3 solutions induced a color change on the surface of the foams, which seemed to be related to the amount of AgNPs present on the foams. Thus, a colorimetric analysis was performed to determine accurately the color coordinates in each foam. Then, this information can be related to the preparation conditions (i.e., reaction time and precursor solution concentration) to check if it allows monitoring the amount of AgNPs synthesized on the foams. Average values and standard deviations of the colorimetric measurements are shown in Table 1.
As a first observation, there is a direct relationship between the coordinate L with the reaction time. In fact, ME foams present the higher values of L* (Luminosity) corresponding to their light gray color. Then, this value decreases considerably by about 30.70 and 35.44% after being immersed for one day in 0.85 and 1.70 mg/mL AgNO 3 solutions, respectively. Moreover, after 7 days of reaction time, the values decrease by about 61.30 and 62.90%. Therefore, the luminosity (L) seems to be quite sensible to the presence of AgNPs, providing noticeable differences for AgNPs contents ranging from about 1 to 18 wt.% regarding the pure ME foams. Thus, it was tested if the luminosity variation (∆L*) term, obtained following Equation (1), might provide information about the amount of AgNPs. First, the obtained ∆L* values from ME/Ag foams obtained after 1, 3, 5, or 7 days of reaction time were related to their known amount of AgNPs (previously determined by ICP-OES elsewhere [10]) (Figure 5a). It was found that these values follow a clear trend, which can be accurately fitted using an exponential expression (R 2 = 0.96). It should be noticed that this expression is not related to the precursor solution or reaction time employed, but only to the amount of AgNPs of the foam. Therefore, it can be applied as a general equation to determine the amount of AgNPs synthesized on the ME foams using both 0.85 and 1.70 mg/mL precursor solutions and any reaction time (Figure 5b).
The obtained values of the AgNPs content (wt.%) from the ∆L* measurements shown in Figure 5b were in good agreement with the known trends for both concentrations along with the entire time range. As a first observation, there is a direct relationship between the coordinate L with the reaction time. In fact, ME foams present the higher values of L * (Luminosity) corresponding to their light gray color. Then, this value decreases considerably by about 30.70 and 35.44% after being immersed for one day in 0.85 and 1.70 mg/mL AgNO3 solutions, respectively. Moreover, after 7 days of reaction time, the values decrease by about 61.30 and 62.90%. Therefore, the luminosity (L) seems to be quite sensible to the presence of AgNPs, providing noticeable differences for AgNPs contents ranging from about 1 to 18 wt.% regarding the pure ME foams.
Thus, it was tested if the luminosity variation (ΔL * ) term, obtained following Equation (1), might provide information about the amount of AgNPs. First, the obtained ΔL * values from ME/Ag foams obtained after 1, 3, 5, or 7 days of reaction time were related to their known amount of AgNPs (previously determined by ICP-OES elsewhere [10]) (Figure 5a). It was found that these values follow a clear trend, which can be accurately fitted using an exponential expression (R 2 = 0.96). It should be noticed that this expression is not related to the precursor solution or reaction time employed, but only to the amount of AgNPs of the foam. Therefore, it can be applied as a general equation to determine the amount of AgNPs synthesized on the ME foams using both 0.85 and 1.70 mg/mL precursor solutions and any reaction time (Figure 5b).
The obtained values of the AgNPs content (wt.%) from the ΔL * measurements shown in Figure  5b were in good agreement with the known trends for both concentrations along with the entire time range. However, the previous approach only considers one of the color coordinates provided by the colorimetry measurements (Table 1). In order to check if the measurement method can be improved  However, the previous approach only considers one of the color coordinates provided by the colorimetry measurements (Table 1). In order to check if the measurement method can be improved Polymers 2020, 12, 996 9 of 15 by incorporating the information of the three color coordinates, the ∆E* ab (Equation (2)) was used following the same procedure and retaking the pure foams as reference. Figure 6a shows the exponential fit (R 2 = 0.97) of the values of ∆E* ab related to the reference values of AgNPs. Yet, it is possible to use the obtained exponential expression to calculate the amount of AgNPs synthesized on the ME foams during the ME/Ag foams preparation (Figure 6b). Once again, the obtained values using this colorimetric approach provided AgNPs content values within the expected trends.
Polymers 2020, 12, 996 9 of 15 by incorporating the information of the three color coordinates, the ΔE * ab (Equation (2)) was used following the same procedure and retaking the pure foams as reference. Figure 6a shows the exponential fit (R 2 = 0.97) of the values of ΔE*ab related to the reference values of AgNPs. Yet, it is possible to use the obtained exponential expression to calculate the amount of AgNPs synthesized on the ME foams during the ME/Ag foams preparation (Figure 6b). Once again, the obtained values using this colorimetric approach provided AgNPs content values within the expected trends. Therefore, it was found that both approaches involving colorimetry measurements provided values for the load of AgNPs in ME foams that comply with the expected trends, independently of the preparation parameters of the foams. Accordingly, this method proved to be reliable for the determination of the amount of AgNPs, with the additional advantages of being a non-invasive, macroscopic, inexpensive, fast, and simple measurement method. Figure 7 shows the acquired radiographs of the ME and the ME/Ag foams after being introduced in two solutions of AgNO3 with different concentrations, 0.85 mg/mL and 1.70 mg/mL, for reaction times ranging from 1 to 7 days (Figure 7a and b respectively). In both figures, a decrease in the foams' X-ray transmissivity is detected with increasing reaction time. Considering that the thickness of the samples was kept practically constant (2.88 ± 0.16 mm) the decrease in X-ray transmissivity can only be attributed to a change in the foams' attenuation coefficient because of the synthesized AgNPs on the surface of the ME foams. Besides, in the radiographs, some features of the cellular structure can be observed, such as the presence of large pores (in the millimetric range). Therefore, it was found that both approaches involving colorimetry measurements provided values for the load of AgNPs in ME foams that comply with the expected trends, independently of the preparation parameters of the foams. Accordingly, this method proved to be reliable for the determination of the amount of AgNPs, with the additional advantages of being a non-invasive, macroscopic, inexpensive, fast, and simple measurement method. Figure 7 shows the acquired radiographs of the ME and the ME/Ag foams after being introduced in two solutions of AgNO 3 with different concentrations, 0.85 mg/mL and 1.70 mg/mL, for reaction times ranging from 1 to 7 days (Figure 7a,b respectively). In both figures, a decrease in the foams' X-ray transmissivity is detected with increasing reaction time. Considering that the thickness of the samples was kept practically constant (2.88 ± 0.16 mm) the decrease in X-ray transmissivity can only be attributed to a change in the foams' attenuation coefficient because of the synthesized AgNPs on the surface of the ME foams. Besides, in the radiographs, some features of the cellular structure can be observed, such as the presence of large pores (in the millimetric range). It is possible to detect how there was nearly total X-ray transmissivity by the pure ME foams prior to its immersion in the AgNO3 solutions. Melamine is a lightweight organic polymer; it is composed of elements with few electrons in their nucleus, such as nitrogen, hydrogen, and carbon. For X-rays, the probability of being attenuated correlates strongly with the number of electrons of an element (i.e., the atomic number Z). Therefore, metals, such as silver, induce strong X-ray attenuation, whereas lighter elements, such as hydrogen, attenuate weakly [41].

X-ray Study of the ME/Ag Foams
In order to test whether the variation in the attenuation coefficient could be used as input to determine the silver content incorporated during the in situ synthesis, a calibration trend is sought by relating the obtained attenuation values to the reference AgNPs values (Figure 8a). A linear relationship was found, revealing that with increasing content of Ag, the attenuation of X-rays by the samples increased too (R 2 = 0.96). In Figure 8b, it is possible to observe the evolution of the AgNPs content as obtained from the attenuation coefficient of the foams. Despite slight deviations in the absolute value, it can be detected how the increase in AgNPs contents determined with the use of X-ray imaging follows the trends observed with ICP-OES. It is possible to detect how there was nearly total X-ray transmissivity by the pure ME foams prior to its immersion in the AgNO 3 solutions. Melamine is a lightweight organic polymer; it is composed of elements with few electrons in their nucleus, such as nitrogen, hydrogen, and carbon. For X-rays, the probability of being attenuated correlates strongly with the number of electrons of an element (i.e., the atomic number Z). Therefore, metals, such as silver, induce strong X-ray attenuation, whereas lighter elements, such as hydrogen, attenuate weakly [41].
In order to test whether the variation in the attenuation coefficient could be used as input to determine the silver content incorporated during the in situ synthesis, a calibration trend is sought by relating the obtained attenuation values to the reference AgNPs values (Figure 8a). A linear relationship was found, revealing that with increasing content of Ag, the attenuation of X-rays by the samples increased too (R 2 = 0.96). In Figure 8b, it is possible to observe the evolution of the AgNPs content as obtained from the attenuation coefficient of the foams. Despite slight deviations in the absolute value, it can be detected how the increase in AgNPs contents determined with the use of X-ray imaging follows the trends observed with ICP-OES.

Accuracy of the Proposed Approaches
AgNPs contents on the obtained ME/Ag foams have been determined using four different characterization methods: a conventional destructive TGA analysis, two approaches based on colorimetry, and one on X-ray radiography. Figure 9 and Table S2 summarize all the obtained results with the different techniques from the set of ME/Ag foams obtained from AgNO3 concentrations of 0.85 and 1.70 mg/mL, (Figure 9a,b, respectively).
(a) (b) Figure 9. AgNPs contents of ME/Ag foams as a function of the reaction time using 0.85 mg/mL (a) and 1.70 mg/mL (b) as solution precursor.
As abovementioned, the three non-invasive approaches accurately reproduce the evolution of the AgNPs content with time previously found by ICP-OES results, independently of the AgNO3 concentration employed [10]. On the contrary, the result provided by TGA, seems to be not accurate for AgNPs contents below 5-8 wt.%.
A more detailed comparative analysis of these techniques is presented in Table 2, where the mean squared errors (Equation (5)) of each technique with respect to the trend of the ICP-OES values, are presented. As expected, the TGA analysis is found to be the less accurate (MSE = 2.05), while the proposed non-invasive approaches present similar values about or below 1. It should be noticed that according to these results the most accurate approach is the measurement of the changes

Accuracy of the Proposed Approaches
AgNPs contents on the obtained ME/Ag foams have been determined using four different characterization methods: a conventional destructive TGA analysis, two approaches based on colorimetry, and one on X-ray radiography. Figure 9 and Table S2 summarize all the obtained results with the different techniques from the set of ME/Ag foams obtained from AgNO 3 concentrations of 0.85 and 1.70 mg/mL, (Figure 9a,b, respectively).

Accuracy of the Proposed Approaches
AgNPs contents on the obtained ME/Ag foams have been determined using four different characterization methods: a conventional destructive TGA analysis, two approaches based on colorimetry, and one on X-ray radiography. Figure 9 and Table S2 summarize all the obtained results with the different techniques from the set of ME/Ag foams obtained from AgNO3 concentrations of 0.85 and 1.70 mg/mL, (Figure 9a,b, respectively).
(a) (b) Figure 9. AgNPs contents of ME/Ag foams as a function of the reaction time using 0.85 mg/mL (a) and 1.70 mg/mL (b) as solution precursor.
As abovementioned, the three non-invasive approaches accurately reproduce the evolution of the AgNPs content with time previously found by ICP-OES results, independently of the AgNO3 concentration employed [10]. On the contrary, the result provided by TGA, seems to be not accurate for AgNPs contents below 5-8 wt.%.
A more detailed comparative analysis of these techniques is presented in Table 2, where the mean squared errors (Equation (5)  As abovementioned, the three non-invasive approaches accurately reproduce the evolution of the AgNPs content with time previously found by ICP-OES results, independently of the AgNO 3 concentration employed [10]. On the contrary, the result provided by TGA, seems to be not accurate for AgNPs contents below 5-8 wt.%.
A more detailed comparative analysis of these techniques is presented in Table 2, where the mean squared errors (Equation (5)) of each technique with respect to the trend of the ICP-OES values, are presented. As expected, the TGA analysis is found to be the less accurate (MSE = 2.05), while the proposed non-invasive approaches present similar values about or below 1. It should be noticed that according to these results the most accurate approach is the measurement of the changes on the luminescence (L*) color coordinate, which is also the more simple, quick, and inexpensive procedure proposed. Table 2. Mean squared error (Equation (5)) of the different values obtained for the AgNPs contents of ME/Ag foams using 0.85 and 1.70 mg/mL as solution precursor, as well as the different reaction times.

Samples
TGA-ICP ∆L-ICP ∆E-ICP X-ray-ICP ME/Ag 2.05 0.31 0.85 1.03 Therefore, the proposed non-invasive characterization approaches based on both colorimetry and X-ray radiography have been proved to be accurate on the determination of the AgNPs content of the ME/Ag foams. As these approaches rely on simple physical principles, it can be expected that their use could be extended to other nanocomposites when the differences between the polymer and the nanofiller are significative in terms of their color or X-ray absorbance. Thus, these techniques are promising approach for the characterization, production monitoring, or quality control of polymer nanocomposites incorporating noble metal nanoparticles (e.g., Au, Pd, Ag), which generally shown characteristic intense colors and present a higher X-ray attenuation than polymers [10,23].

Conclusions
This work developed novel non-invasive procedures for the study and characterization of the filler load on polymer nanocomposites. In particular, two approaches based on colorimetric analysis and one on X-ray radiography were successfully tested on a wide set of ME/Ag foams obtained by the in situ synthesis of Ag NPs on melamine (ME) foams.
On the one hand, the colorimetry-based techniques take advantage of the color change induced on the ME/Ag foams by the presence of AgNPs, which turn the light gray ME foams into brown tonalities that darkens with increasing amounts of AgNPs. A direct relationship between the luminosity and the overall color change of a reference set of ME/Ag samples and their known AgNPs load was found, allowing to obtain calibration equations relating these values. From these equations, it was possible to determine the load of the entire set of ME/Ag foams, finding an excellent agreement between those values and the known trends for the reference ME/Ag foams.
On the other hand, the differences on the X-ray absorption coefficients between the polymer matrix and the filler, in this case AgNPs, also allows the quantification of the AgNPs load by X-ray radiography. It was possible to calculate the absorption coefficients of the reference set of ME/Ag foams and relate them to their AgNPs load, obtaining again a calibration equation which accurately reproduce the expected trends for the entire set of ME/Ag samples.
In addition, a conventional destructive technique such as TGA was employed for comparison purposes. It was found that for the ME/Ag samples this technique provides less accurate results than those obtained with the non-invasive techniques proposed, as the presence of ME residue even at very-high temperatures, and the possible interaction between the ME and AgNPs during the ME degradation, make this method inaccurate for low AgNPs contents.
Therefore, the proposed non-invasive procedures proved to be suitable and facile approaches for the determination of the AgNPs load on the ME/Ag nanocomposite foams. Moreover, the colorimetry-based approaches proved to be the most accurate, while this technique is also inexpensive and can be easily carried out in situ. Also, these techniques could be applied to other nanocomposites with proper features in terms of color or X-ray attenuation differences.