Pyrolysis of chitosan (CAS 9012-76-4, high molecular weight 310–375kDa, DD=75–85%, Aldrich) was performed in a quartz tube reactor (Carbolite MTF 12/38/400) under Ar atmosphere with the following thermal program: ramp from room temperature (RT) to target temperature (either 440°C or 800°C, rate: 5°Cmin−1), isothermal step 1h. After the pyrolysis treatment, samples were washed twice with deionized water to remove all soluble residues, and dried overnight at 80°C in an oven. Chars produced were coded as follows: C440 refers to char obtained after heating at 440°C, whereas C800 refers to the one heated at 800°C. Samples preparation is reported in Table 1.

An American Petroleum Institute (API) oil-well cement Class G (Lafarge North America) was used in this work.

Either chitosan or its char (only C800) were dispersed in deionized water (0.1% by weight of cement, or BWOC as in [17]) with an ultrasonic tip (Vibra-cell™) for 15min at 100W. The suspensions were mechanical stirred while cement powder was added into the aqueous medium. The water-to-cement (W/C) ratio selected was 0.45. In this work, dispersant or surfactant agents or other additives were not used. The slurry, composed only by cement and chitosan/char, was then poured into molds. The curing for this category of cement is 24h at 85°C and at 100% relative humidity. Prismatic molds of size 20×20×80mm were used for building the cement-based composites. Composites were coded as follows: CC refers to the cement-chitosan composite, whereas CC800 refers to the cement composite with char C800. Additionally, for the sake of comparison, also bare cement samples were prepared and identified with the acronym C0.

Thermo-gravimetric analysis was conducted in a TGA instrument Mettler Toledo 1600, under Ar atmosphere. The analysis was performed on pristine precursor (chitosan) with the following thermal program: heating ramp from 25°C to 1000°C (10°Cmin−1). The Ar was supplied with a constant flow rate (50mLmin−1).

Field emission scanning electron microscopy (FESEM) micrographs were collected by means of a FESEM Zeiss Merlin. Various micrographs were taken in different areas of the samples analyzed in order to observe the average morphology.

Elemental analysis of precursor and char was carried out in a Thermo FlashEA 1112 CHNS-O analyzer. Average values of three replicas were provided for each measurement.

FTIR spectra were recorded in attenuated total reflectance mode (ATR, equipped with diamond cell for single reflection) by means of a Vector 22 spectrometer (Bruker) equipped with Globar source, DTGS detector and working with 128 scans at 4cm−1 of resolution in the 4000–400cm−1 range. Spectra were recorded directly on powdery samples.

Mechanical properties were measured with a three-point bending test using the procedure prescribed by JCI-S-001-2003. The specimens were tested with a single-column Zwick-Line z050 flexural testing machine with a load cell having a maximum capacity of 1kN. As described in the literature [43], the tests were performed on notched specimens (notch width: 2mm, depth: 5mm, i.e., equal to one fourth of the width of specimen) by controlling the crack mouth opening displacement (CMOD). CMOD was controlled at a fixed displacement rate of 0.005mmmin−1 by placing an extensometer on the two sides of the notch. All tests were performed at least on three specimens, and average values were provided.

The flexural strength was calculated by using the standard formula for un-notched bending tests, using the thickness in the notched area. The toughness was calculated by integrating the load-displacement curve as in [43]. Compressive strength tests were conducted on at least three cubic specimens for each sample (size: 20×20×20mm) and controlled at a fixed displacement rate of 0.01mm min−1.

To evaluate the pyrolysis induced effect on chitosan structure, TGA analysis was performed on chitosan under inert/reducing atmosphere. Fig. 1 reports the TG profile, which presents two main weight losses: the first one at ca. 100°C due to the loss of physisorbed water molecules (7wt.%, i.e., chitosan is a hygroscopic material) and the second one in the 250–400°C range due to the degradation of the polysaccharidic structure with formation of volatiles products [44]. According to the literature [39], pyrolysis of chitosan leads to the formation of H2O, CO2, CO, NH3, CH3COOH and N-containing heteroaromatic compounds (mostly pyrazines, pyridines and pyrroles). The curve profile evidences also that at higher temperature (>450°C) the degradation still continues very slowly, with release of CH4 and consequent formation of a carbonaceous residue (ca. 27wt.% at 1000°C) via dehydrogenation mechanism [39].

Fig. 1.

TG (black solid) and DTG (blue dotted) analyses of chitosan heated under Ar atmosphere.

(0.1MB).

For a more detailed comprehension of the chitosan-to-char conversion mechanisms, elemental analysis was performed on either bare chitosan or its chars, sampling at two different target temperatures: namely 440°C (before) and 800°C (after the formation of methane documented by the literature for analogous systems) [39]. The analysis revealed a progressive increment of the content of C (from 41.2% to 72.9%) and a reduction of H (from 6.6% to 1.3%). Interestingly, the amount of N increases from pristine chitosan (7.3%) to C440 char (9.8%), while it subsequently decreases to 6.8% for C800. The explanation of this trend is assumable to the fate of N during chitosan pyrolysis. In fact, it has been demonstrated that the amino groups of chitosan can evolve toward either volatile ammonia or being integrated into aromatic structures. Such structures can evolve into a solid N-containing residue (N-doped char) as well as into volatiles N-containing heteroaromatic compounds [39,45]. Therefore, one can easily assume that at high temperatures (800°C), the loss of residual functionalities is remarkably higher than at low temperature (440°C). This trend is in agreement with previous study devoted to the pyrolysis of chitin [40] (Table 2).

To deepen the conversion mechanism, ATR–FTIR spectra on chitosan and its chars were collected and results shown in Fig. 2. As evidenced in the figure, the IR spectrum of pristine chitosan revealed the presence of an intense and broad band centered at ca. 3440cm−1 attributable to the stretching mode of O–H and N–H functionalities, followed by a small signal in the 2850–2880cm−1 due to stretching mode of aliphatic C-H, the two characteristic signals at ca. 1655cm−1 due to the axial stretching vibration mode of acetamido group (amide I) and the one at ca. 1580cm−1 due to the angular deformation of amino groups, the signals in the 1300–1500cm−1 range commonly assigned to C–N stretching and N–H deformation as well as CH3 deformation modes, and lastly a broad signal centered at 1000cm−1 due to the skeletal C–O and C–O–C stretching modes. Since chitosan is hydroscopic, the signals at ca. 3440cm−1 and at ca. 1655cm−1 could be associated also to the presence of physisorbed water molecules (O–H stretching and bending modes, respectively). Pyrolysis carried out at high temperatures (both 440°C and 800°C) caused the loss of the characteristic glycosidic functional groups with growth of a broad signal in the 1600–1000cm−1 attributable to the formation of carbonaceous CC and C–C stretching modes, with presence of an IR signal at ca. 3200–3400cm−1 attributable to N–H residual polar functionalities (or still sorbed water molecules) [40]. Additionally, it should be notice that the aliphatic C–H stretching modes at 2850–2880cm−1 are still present in traces in C440, whereas they are absent in the C800 (in agreement with the elemental analysis).

Fig. 2.

ATR–FTIR spectra in the 4000–400cm−1 range of chitosan (green curve), and its chars C440 (blue curve) and C800 (red curve).

(0.1MB).

Fig. 3 reports the main morphological differences between precursor prior to perform pyrolysis (Fig. 3A) and at the end of the pyrolysis process (C800, Fig. 3B). Images revealed a rather dense corrugate surface in the case of bare chitosan. On the contrary, pyrolysis treatment favored the formation of surfaces characterized by the presence of debris and defects. The analysis of the cross-section evidenced also the formation of cracks and delamination sites.

Fig. 3.

FESEM micrographs of bare chitosan (A) and its char C800 (B). Insets refer to the same materials in cross-section at low magnification.

(0.2MB).

Fig. 4 and Table 3 report the results of mechanical tests performed on bare cement (C0) and on composites containing either chitosan (CC) or chitosan-derived char (CC800). Results demonstrate that composites with chitosan possess lower flexural strength compared to reference cement, whereas both toughness and compression strength are not-significantly different between each other. C800-containing materials, instead, show a slight reduction of both flexural strength and compression strength, whereas the relative toughness value is slightly higher than both bare cement and chitosan-containing composites. The loss in terms of flexural strength registered in both cases could be associated to a certain inhomogeneity of samples caused by a difficult interaction between chitosan and its char with the hydrophilic cement paste. On the other hand, the dispersion of the fillers in the cement paste assumes a key role. A bad dispersion of the filler can create defect in the cement matrix composite with a worsening in the mechanical resistance. The presence of functional groups in bare chitosan should favor its integration within the cement matrix, although it must be notice that chitosan is hygroscopic and might compete with the hydration processes involved during the setting step. Moreover, interfacial phenomena are surely the main concern for char-containing composites since char is chemically reduced (and hydrophobic). However, all these issues could be overcome by introducing in the formulation dispersants and compatibilizers (e.g., amphiphilic molecules) or further additives, as well as the chemical functionalization of the char surface.

Fig. 4.

Mechanical properties of cementitious composites. Panel A: load–displacement curves for flexural strength. Panel B: mechanical properties (flexural strength, toughness, and compression strength) of C0 (black), CC (green), and CC800 (red).

(0.14MB).

In general, the analysis of these data (even if not rousing) are still promising since highlights the possibility of successfully introduce chitosan or its char without causing a drastic depletion of the mechanical performances in cement. Furthermore, it must be notice that the experimental value for the toughness in char containing composites is very encouraging. It can be note that the improvement of the toughness response is attributable to the bridging effect of the particles of char dispersed in the cement matrix. In analogy with graphene (see for instance [46,47]), this effect allows a delay in the propagation of micro-cracks if the composite is subjected to an applied stress. The load vs. displacement profile for CC800 confirmed this behavior, demonstrating an incipient fracture toughness capability. Hence, even if these results are only preliminary (and deserve a significant improvement), experiments revealed being an interesting starting point for the development of sustainable toughened cementitious composites.