**3. Serinol derivative for graphene functionalization**

#### **3.1. Serinol derivative: 2‐(2,5‐dimethyl‐1H‐pyrrol‐1‐yl)‐1,3‐propanediol**

To pursue the (i)–(iii) objectives reported in Introduction, a serinol derivative was used: 2‐ (2,5‐dimethyl‐1H‐pyrrol‐1‐yl)‐1,3‐propanediol, in the text referred to as serinol pyrrole (SP). Chemical structure of serinol pyrrole is shown in **Figure 1**.

In this paragraph, synthesis of serinol pyrrole and motivation for using such a molecule in combination with a high surface area nanosized graphite (HSAG) is discussed.

Serinol is 2‐amino‐1,3‐propane diol that can be directly obtained from renewable sources [60] and is produced at the industrial scale from glycerol. Glycerol is a non‐toxic and biodegradable raw material, cheap and easily available as it is the main co‐product of biodiesel production.

**Figure 1.** 2‐(2,5‐dimethyl‐1*H*‐pyrrol‐1‐yl)‐1,3‐propanediol (serinol pyrrole, SP).

Glycerol can be thus considered as a waste of an important industrial process. To use and to give added value to such a waste is particularly meaningful, as glycerol can be the building block for a C3 platform alternative to the oil‐based one. It is well known that large research efforts are made nowadays to replace oil with biomass as the source for preparing chemi‐ cals and materials. First and second generation biomasses are identified. The first generation includes biomass that has impact on the food chain. For example, first‐generation bioethanol is produced from sucrose, oligosaccharides and starch, whereas lignocellulose materials are used for the second‐generation bioethanol. Biodiesel was produced from oil crops, whereas nowadays it is increasingly obtained from algae. In order to reduce the use of biomasses that could be employed for food production, the exploitation of wastes is encouraged. The European Parliament has made clear statements about the use of wastes and residues in place of original biomass: 'The recent experience of the development of certain renewable energy sources, particularly biofuels from food and feed crops such as cereals, oilseeds and sugar, has stimulated concern that new biorefinery processes must as far as possible be based on non‐competing wastes and residues to minimize impacts on food availability and prices' [61]. More than 1 million tons of glycerol are produced per year, with a price well below 1 Euro/kg, and many routes are available for the selective conversion of glycerol to value‐added prod‐ ucts [62–65]. Serinol, produced via its reductive amination, presents the precious tool of che‐ moselectivity between amino and hydroxyl groups which allows in developing innovative synthetic strategies. Indeed, objective of our research was to exploit the chemoselectivity of serinol, transforming the amino moiety into a group suitable for interaction with sp2 carbon allotropes and, in particular, with graphene layers.

#### **3.2. Synthesis of serinol derivative, 2‐(2,5‐dimethyl‐1H‐pyrrol‐1‐yl)‐1,3‐propanediol**

#### *3.2.1. Neat synthesis*

It is evident that the objective of developing a really simple method, suitable for large‐scale

The reduction step is crucial in order to obtain graphene. Well‐known methods are based on the use of hydrazine [47–50] or hydrogen plasma [49]. To avoid reaggregation of graphitic flakes, a stabilizing agent should be added [9, 51]. Hydrazine is well known as a toxic reagent. Hence, eco‐friendly methods have been attempted. Quick deoxygenation of graphite oxide assisted by a base (NaOH, 0.1 M) has been performed at moderate temperatures (80°C) [52]. Ascorbic acid has been used as well [53]. Thermal [54] and flash [55] reductions have been

However, it is widely acknowledged that reduction is still incomplete and that the ideal graphene structure is neither preserved nor restored [20, 56]. Hence, also this objective is not

However, the pathway that could lead to graphene through graphite oxidation gives the opportunity of preparing GO that could be a suitable building block for further reactions [24–26] as well as for catalytic applications [17, 18]. Moreover, functionalized single‐layer graphene sheets can be prepared by splitting graphite oxide [57]. The structure of GO has been investigated for decades, but it is substantially still unknown [20, 21, 58], and this makes GO not the ideal building block for further reactions. It was reported [59] that hydroxyl and epoxide groups are on the surface of basal planes and carbonyl and carboxyl groups are on the edges. Moreover, it is widely acknowledged that oxidation leads to extensive disruption

hybridization of graphene layers, that is, graphene properties are drastically damaged

This brief sum up allows to comment that important research efforts are required in order to achieve the objective of preparing graphene and few‐layer graphene (optionally containing functional groups) preserving the ideal graphene structure and through a simple, eco‐friendly,

To pursue the (i)–(iii) objectives reported in Introduction, a serinol derivative was used: 2‐ (2,5‐dimethyl‐1H‐pyrrol‐1‐yl)‐1,3‐propanediol, in the text referred to as serinol pyrrole (SP).

In this paragraph, synthesis of serinol pyrrole and motivation for using such a molecule in

Serinol is 2‐amino‐1,3‐propane diol that can be directly obtained from renewable sources [60] and is produced at the industrial scale from glycerol. Glycerol is a non‐toxic and biodegradable raw material, cheap and easily available as it is the main co‐product of biodiesel production.

production, has not been achieved yet.

178 Graphene Materials - Structure, Properties and Modifications

and GO loses the benefits of a graphitic structure.

**3. Serinol derivative for graphene functionalization**

Chemical structure of serinol pyrrole is shown in **Figure 1**.

**3.1. Serinol derivative: 2‐(2,5‐dimethyl‐1H‐pyrrol‐1‐yl)‐1,3‐propanediol**

combination with a high surface area nanosized graphite (HSAG) is discussed.

economically viable and scalable method.

reported.

achieved yet.

of sp2

As it is shown in **Scheme 1**, Paal‐Knorr reaction [66, 67] was carried out between 2‐amino‐1,3‐ propane diol (serinol, S) and 2,5‐hexanedione (HD), transforming the amino group into a pyr‐ role ring and thus giving rise to 2‐(2,5‐dimethyl‐1H‐pyrrol‐1‐yl)‐1,3‐propanediol (SP) [68–71].

**Scheme 1.** Neat reaction between serinol and 2,5‐hexanedione for preparing serinol pyrrole. S and HD in equimolar amount.

It is definitely worth observing that the reaction that leads to SP is characterized by very high atom economy, 82.5%, with H2 O as the only by‐product. Neat reaction was carried out, in the absence of solvents and catalysts [68], with temperatures ranging from 130 to 180°C and reaction times rang‐ ing from 30 to 120 min. In Section 8, an example characterized by very high reaction yield to SP is reported: 96%. In consideration of this yield and of the atom economy, the atom efficiency is 85%.

**Scheme 1** reveals a two‐step synthesis. The first step is the reaction between equimolar amounts of S and HD and leads to the preparation of a tricyclic compound: 4a,6a‐dimethyl‐hexahydro‐1,4‐ dioxa‐6b‐azacyclopenta[cd]pentalene (HHP). The reaction was performed at room temperature, with almost complete conversion and selectivity, the only by‐product being water (as mentioned above). In the second step of the reaction, the tricyclic compound was then isomerized to the aromatic SP, with high conversion and selectivity, by simply increasing the reaction tempera‐ ture. Indeed, Nuclear Magnetic Resonance (NMR) spectra of the products of reaction steps 1 and 2 did not reveal the presence of any chemical compounds other than HHP and SP, respec‐ tively. The Paal‐Knorr reaction between S and HD had been already reported [72]. Reagents were refluxed in toluene in the presence of acidic substances such as glacial acetic acid (in large amount) and *p‐*toluenesulfonic acid (in catalytic amount). The product mixture contained SP in low amount (yield was about 13% mol), the tricyclic compound (HHP) and polymeric materi‐ als, whose relative amount was enhanced by acids. The neat two‐step synthesis described here surprisingly led to SP with high atom efficiency. As published by some of the authors [69], this result could be ascribed, at least in part, to absence of acids and mild reaction conditions. Such experimental frame favoured, in the first reaction step, hemiaminal formation and polycycliza‐ tion, preventing aromatization to SP and formation of polymeric species. Temperature increase allowed then the establishment of the thermodynamic control of the reaction.

#### *3.2.2. Synthesis on high surface area nanographite (HSAG)*

Synthesis of serinol pyrrole was as well performed on high surface area nanosized graphite, HSAG. Characteristics of HSAG are discussed in Section 4.1. The simple reaction process is shown in **Scheme 2**.

Details are in Section 8. Nuclear Magnetic Resonance (NMR) analysis of the solution of the reaction product in D2 O revealed the presence of serinol pyrrole only. This result indicates that the reaction between S and HD occurs with high yields also by supporting the ingredi‐ ents on a graphitic substrate.

**Scheme 2.** Reaction between serinol (S) and 2,5‐hexanedione (HD) on high surface area graphite (HSAG), for the selective preparation of serinol pyrrole (SP). S and HD are in equimolar amount.

#### **3.3. Why to use serinol pyrrole for the functionalization of graphene layers?**

It is definitely worth observing that the reaction that leads to SP is characterized by very high atom

**Scheme 1.** Neat reaction between serinol and 2,5‐hexanedione for preparing serinol pyrrole. S and HD in equimolar

solvents and catalysts [68], with temperatures ranging from 130 to 180°C and reaction times rang‐ ing from 30 to 120 min. In Section 8, an example characterized by very high reaction yield to SP is reported: 96%. In consideration of this yield and of the atom economy, the atom efficiency is 85%. **Scheme 1** reveals a two‐step synthesis. The first step is the reaction between equimolar amounts of S and HD and leads to the preparation of a tricyclic compound: 4a,6a‐dimethyl‐hexahydro‐1,4‐ dioxa‐6b‐azacyclopenta[cd]pentalene (HHP). The reaction was performed at room temperature, with almost complete conversion and selectivity, the only by‐product being water (as mentioned above). In the second step of the reaction, the tricyclic compound was then isomerized to the aromatic SP, with high conversion and selectivity, by simply increasing the reaction tempera‐ ture. Indeed, Nuclear Magnetic Resonance (NMR) spectra of the products of reaction steps 1 and 2 did not reveal the presence of any chemical compounds other than HHP and SP, respec‐ tively. The Paal‐Knorr reaction between S and HD had been already reported [72]. Reagents were refluxed in toluene in the presence of acidic substances such as glacial acetic acid (in large amount) and *p‐*toluenesulfonic acid (in catalytic amount). The product mixture contained SP in low amount (yield was about 13% mol), the tricyclic compound (HHP) and polymeric materi‐ als, whose relative amount was enhanced by acids. The neat two‐step synthesis described here surprisingly led to SP with high atom efficiency. As published by some of the authors [69], this result could be ascribed, at least in part, to absence of acids and mild reaction conditions. Such experimental frame favoured, in the first reaction step, hemiaminal formation and polycycliza‐ tion, preventing aromatization to SP and formation of polymeric species. Temperature increase

allowed then the establishment of the thermodynamic control of the reaction.

Synthesis of serinol pyrrole was as well performed on high surface area nanosized graphite, HSAG. Characteristics of HSAG are discussed in Section 4.1. The simple reaction process is

Details are in Section 8. Nuclear Magnetic Resonance (NMR) analysis of the solution of the

that the reaction between S and HD occurs with high yields also by supporting the ingredi‐

O revealed the presence of serinol pyrrole only. This result indicates

*3.2.2. Synthesis on high surface area nanographite (HSAG)*

O as the only by‐product. Neat reaction was carried out, in the absence of

economy, 82.5%, with H2

180 Graphene Materials - Structure, Properties and Modifications

amount.

shown in **Scheme 2**.

reaction product in D2

ents on a graphitic substrate.

SP is a serinol derivative, hence it comes from a molecule that is naturally occurring whether it comes from natural sources or from a waste such as glycerol. In the light of that, SP can be considered as a bio‐sourced molecule. Its preparation appears to have also another good fea‐ ture in view of a sustainable process, that is, the high atom efficiency.

By observing the chemical structure of serinol pyrrole in **Figure 1**, it can be commented that SP is a *Janus* molecule [70], that is, a molecule with two faces. Such a definition, that comes from the portray of a Roman god, was first used to indicate micro‐ and nanoparticles with at least two physically or chemically different surfaces [73, 74] and, subsequently, to describe molecules with two moieties (one hydrophobic and one hydrophilic) such as colloids [75] and block copolymers [76]. A *Janus* molecule is known to have a dual reactivity.

What are the two moieties in SP? One is represented by the pyrrole ring. The Paal‐Knorr reac‐ tion changes the hybridization of the sp3 nitrogen atom of the amino group and leads to the formation of sp2 atoms in the aromatic pyrrole ring, which could give rise to π‐π stacking with aromatic compounds such as graphene layers and hence to a stable interaction; moreover, the pyrrole ring could interact with lipophilic substances. Conversely, the moiety that contains hydroxy groups is hydrophilic and can promote the interaction with polar environments. Moreover, hydroxy groups can be reactive functional groups for the preparation of step growth polymers [70, 72].

Thanks to its hydrophilic moiety, SP could then promote the dispersion of graphitic aggregates in polar solvents, such as water. Exfoliation of such aggregates into few or single graphene lay‐ ers could be thus promoted. As mentioned in the introduction, dilution in appropriate solvents is a method for the preparation of graphene and few‐layer graphene. However, in most cases, organic (mainly aromatic) substances have to be used [22]. The structure of the graphitic sub‐ strate is expected to remain substantially unaltered by the interaction with SP: this would lead to introduce functional groups with heteroatoms such as oxygen and nitrogen on graphene layers that would maintain their pristine structure. It is evident that the success of such a design essentially depends on the strength of the interaction between SP and the graphene lay‐ ers and also on the amount of SP which should be used. Moreover, the choice of the graphitic substrate for the preparation of adducts with SP appears to be of great importance.

### **4. Adducts of graphene layers with serinol pyrrole**

In this paragraph, characteristics of HSAG (selected as the graphitic substrate) preparation and characterization of HSAG‐SP adducts are discussed.

#### **4.1. Graphitic substrate: high surface area nanosized graphite**

High surface area nanosized graphite was selected for the preparation of adducts with SP. Characterization of HSAG has been already reported [57, 70, 77]. Some relevant data are shown in **Table 1**, in comparison with those of other sp2 carbon allotropes: multiwalled car‐ bon nanotubes (CNT) and a furnace carbon black, CBN326.

HSAG has a high surface area and is able to establish extended interactions with a polymer matrix, as shown by the diisobutyl phthalate (DBP) absorption number. By comparing HSAG and CNT data in **Table 1**, a correlation cannot be seen between surface area and DBP absorp‐ tion number. As commented in previous publications [77], this finding could be explained considering that graphene layers in HSAG are stacked in crystalline domains: layers are more accessible to small nitrogen molecules used in Brunauer Emmet Teller (BET) analysis than to bulky phthalates. However, the peak shape analysis of wide angle X‐ray diffraction (WAXD) pattern (shown in **Figure 2**, with the peak assignment), performed by applying the Scherrer equation to (002) reflection [56], revealed quite a low number of graphene layers stacked in a crystalline domain (about 35). Moreover, the analysis of (100) and (110) reflections, typical of the crystalline order inside the layer, allowed to estimate the shape anisotropy, defined as the ratio between the crystallites dimensions in directions parallel and orthogonal to structural lay‐ ers [56]. HSAG was found to have the largest shape anisotropy, when compared with CBN326, expanded graphites and coke and calcinated petroleum cokes. (Transmission electron micros‐ copy (TEM) analyses revealed that HSAG layers had an average size of about 300 nm [70].

Characterization of HSAG was also performed via Raman spectroscopy [70], a crucial tech‐ nique for the study of carbonaceous materials [79–83]. Raman spectrum of HSAG is reported in **Figure 2** below in the text. In the Raman spectrum, two lines, named D and G, reveal the presence of graphitic sp2 ‐phase: they are located at 1350 cm−1 and 1590 cm−1, respectively. Bulk crystalline graphite (graphene) gives rise to G peak, whereas the D peak occurs in the presence of either structural defects or confinement (e.g. by edges) of the graphitic layers [78, 79, 82–84]. Carbon atoms with different hybridization and grafted functional groups are indeed structural defects. Graphene layers have finite dimensions and irregular boundaries whose relative importance depends on the size of the layer. These boundaries contribute to


a From BET measurements (see Ref. [78]).

b mL of absorbed DBP/100 g of CB (see Ref. [78]).

c Estimated from WAXD pattern (see Ref. [56]).

dBaytubes C150 P from Bayer Material Science (see Ref. [78]).

e From Cabot.

**Table 1.** Surface area, DBP absorption number and number of graphene layers stacked in crystalline domain for sp2 carbon allotropes.

**4.1. Graphitic substrate: high surface area nanosized graphite**

shown in **Table 1**, in comparison with those of other sp2

182 Graphene Materials - Structure, Properties and Modifications

presence of graphitic sp2

**Carbon allotrope Surface areaa**

From BET measurements (see Ref. [78]).

mL of absorbed DBP/100 g of CB (see Ref. [78]).

dBaytubes C150 P from Bayer Material Science (see Ref. [78]).

Estimated from WAXD pattern (see Ref. [56]).

a

b

c

e

From Cabot.

carbon allotropes.

 **(m2**

HSAG 330 162 35 CNTd 200 316 10 CB N326e 77 85 5

bon nanotubes (CNT) and a furnace carbon black, CBN326.

High surface area nanosized graphite was selected for the preparation of adducts with SP. Characterization of HSAG has been already reported [57, 70, 77]. Some relevant data are

HSAG has a high surface area and is able to establish extended interactions with a polymer matrix, as shown by the diisobutyl phthalate (DBP) absorption number. By comparing HSAG and CNT data in **Table 1**, a correlation cannot be seen between surface area and DBP absorp‐ tion number. As commented in previous publications [77], this finding could be explained considering that graphene layers in HSAG are stacked in crystalline domains: layers are more accessible to small nitrogen molecules used in Brunauer Emmet Teller (BET) analysis than to bulky phthalates. However, the peak shape analysis of wide angle X‐ray diffraction (WAXD) pattern (shown in **Figure 2**, with the peak assignment), performed by applying the Scherrer equation to (002) reflection [56], revealed quite a low number of graphene layers stacked in a crystalline domain (about 35). Moreover, the analysis of (100) and (110) reflections, typical of the crystalline order inside the layer, allowed to estimate the shape anisotropy, defined as the ratio between the crystallites dimensions in directions parallel and orthogonal to structural lay‐ ers [56]. HSAG was found to have the largest shape anisotropy, when compared with CBN326, expanded graphites and coke and calcinated petroleum cokes. (Transmission electron micros‐ copy (TEM) analyses revealed that HSAG layers had an average size of about 300 nm [70].

Characterization of HSAG was also performed via Raman spectroscopy [70], a crucial tech‐ nique for the study of carbonaceous materials [79–83]. Raman spectrum of HSAG is reported in **Figure 2** below in the text. In the Raman spectrum, two lines, named D and G, reveal the

Bulk crystalline graphite (graphene) gives rise to G peak, whereas the D peak occurs in the presence of either structural defects or confinement (e.g. by edges) of the graphitic layers [78, 79, 82–84]. Carbon atoms with different hybridization and grafted functional groups are indeed structural defects. Graphene layers have finite dimensions and irregular boundaries whose relative importance depends on the size of the layer. These boundaries contribute to

**Table 1.** Surface area, DBP absorption number and number of graphene layers stacked in crystalline domain for sp2

‐phase: they are located at 1350 cm−1 and 1590 cm−1, respectively.

**/g) DBP absorption number (mL/100 g)b Number of stacked layersc**

carbon allotropes: multiwalled car‐

**Figure 2.** WAXD patterns (on the left) and Raman spectra (on the right) of HSAG (a), HSAG‐SP‐M (b) and HSAG‐SP‐T (c).

the D band. In the Raman spectrum of HSAG in **Figure 2**, both D and G bands are present, with similar intensities. To justify the intensity of the D band, some of the authors reported an interpretation [70] based on the existence of a confined crown region close to the edge, affected by confinement effects and electronically perturbed. This interpretation takes into account the small size of the HSAG graphitic layers and the obtainment of HSAG through ball milling. It was demonstrated [85] that the intensity of the D band increases in a gra‐ phitic sample by reducing the size of the layers by progressive ball milling.

HSAG appears thus an ideal substrate for the preparation of adducts with SP, thanks to the following features: high surface area, low number of stacked layers, high crystalline order inside the layers and remarkable presence of edges, that could favour interactions or even reactions with a suitable molecule.

#### **4.2. Adducts of SP with HSAG: preparation and characterization**

Adducts were prepared as illustrated in **Scheme 3**. In a nutshell, HSAG and SP were mixed in acetone, the mixture was sonicated for few minutes, solvent was removed and either mechanical or thermal energy was given to the solid HSAG‐SP mixture, via ball milling or simply heating. The following adducts were prepared: HSAG‐SP‐M from mechanical treatment and HSAG‐SP‐T from thermal treatment. Typical examples of the preparation of mechanical and thermal adducts are in Section 8.

The strength of interaction between HSAG and SP was investigated. HSAG‐SP adducts, upon extraction (see Section 8) were weighed and analysed through thermogravimetric analysis (TGA). As reported in a previous publication [70], the amount of SP in the adduct was esti‐ mated on the basis of the mass loss in the temperature range from 150 to 500°C. Data are col‐ lected in **Table 2**. The SP: HSAG ratio was estimated as molar ratio (the moles of HSAG are the moles of benzene rings), while the yield of functionalization was calculated by applying the following equation:

**Scheme 3.** Block diagram for the preparation of HSAG‐SP adducts, by using either mechanical or thermal energy (for details, see Section 8).


a Evaluated from TGA, with the equation reported in the text.

b Variation coefficient, calculated for a population of 10 samples, was 1.7%.

**Table 2.** Preparation of HSAG‐SP adducts: yield of functionalizationa .

\*\*Table 2\*\*.\*\* Perparation of HSAG-SP adults: yield of functionalization!.

$$\text{Yield} = \frac{\text{SP in (HSAG SP addcut)}\_{\text{atm the continuum}}}{\text{SP in (HSAG SP addcut)}\_{\text{hour the reaction}}} \times 100\tag{1}$$

Very high yield of functionalization (larger than 90%) was achieved, both through mechanical and thermal treatments. In particular, thermal treatment with 0.1 as the SP:HSAG molar ratio gave a yield of 99%. This indicates that SP is able to establish strong interactions with HSAG. It is worth noting that the yield of functionalization decreased with the starting SP:HSAG molar ratio.

WAXD and Raman analyses were used to study the organization at the solid state of HSAG‐SP adducts (details are in Ref. [71]). X‐ray diffraction patterns and Raman spectra of HSAG (a), HSAG‐SP‐M (b) and HSAG‐SP‐T (c) are shown in **Figure 2**.

As regards WAXD patterns, in HSAG‐SP‐M and HSAG‐SP‐T, (002) reflection remains at the same 2θ value as in pristine HSAG, indicating that SP was not intercalated in the graphitic interlayer space. The number of layers stacked in a crystalline domain, calculated by applying the Scherrer equation to (002) reflection, as explained above, was found to decrease from HSAG (35) to HSAG‐SP‐T (29) to HSAG‐SP‐M (24). Peak shape analysis was as well performed on (100) and (110) reflections, analysing pristine HSAG and the adducts: the correlation length was found in a range from 26.5 to 28 nm. Hence, the in‐plane order of HSAG was not substantially altered by the reaction with SP, even via ball milling.

Considering Raman characterization, D band is present in the spectrum of pristine HSAG and its relative intensity with respect to G band does not substantially change in HSAG‐SP‐T, after the thermal treatment. This finding suggests that the treatment with SP does not appre‐ ciably modify the structure of HSAG. On the contrary, the intensity of D band increases in HSAG‐SP‐M, after the mechanical treatment. In the light of what observed from WAXD analysis, the increase of disorder in the milled sample could be attributed to electronic per‐ turbation of the confined crown region close to the edge, in line with what reported in the literature [85] about the effect of milling on graphitic samples. Different types of disorder can occur at the very edges: grafting of molecules and loss of sp2 hybridization are among them.

The chemical nature of HSAG‐SP adduct was investigated by means of FTIR analysis (experimental details are in reference [71]). The IR spectra of HSAG, HSAG‐SP‐T, HSAG‐SP‐M and SP, in the region 4000–700 cm−1 are reported in **Figure 3**.

The spectrum of HSAG (a) is characterized by the peak near 1590 cm−1 which can be assigned to the absorption of E1u IR active mode of collective C=C stretching vibration of graphite/ graphene materials. The increasing background towards high wavenumbers is due to diffusion/reflection phenomena of the IR beam passing through HSAG microparticles.

In the SP spectrum (d), the broad band at 3370 cm−1 can be reasonably attributed to hydrogen‐ bonded OH groups. The pyrrole ring is evidenced by the collective vibration mode of C=C/C‐C stretching, located at about: 1530, 1395, 1490 and 802 cm−1. Such bands can be taken as finger‐ print of SP.

Spectra of extracted HSAG‐SP‐T and HSAG‐SP‐M adducts (**Figure 3b** and **c**) show bands that cannot be attributed to HSAG. Such bands are located as follows: at about 2900 cm−1, in the region of sp3 C‐H stretching; at 1590 and 1470 cm−1, in the region of C‐C stretching of aromatic rings; at 1383 cm−1, in the region of vibrations of diols and at 802 cm−1, where vibration of the alkenyl groups absorbs. C‐H stretching can be due to methyl groups of the pyrrole ring and to methylenes of the serinol moiety. Diols come from serinol as well. Aromatic C‐C stretching is due to the presence of pyrrole ring. Moreover, in both the adducts' spectra, there are bands that are not present in either HSAG or SP spectra. Such bands are located at 1742 and 1662 cm−1. Absorbances in this region are usually due to the presence of carbonyl groups. In particular, these new bands can be assigned to an aldehydic group. As reported in Section 8, the preparation of adducts is performed at high temperatures and in the presence of air. Oxidation process of the pyrrole ring could be thus hypothesized. Moreover, the efficient synthesis of graphene sheets using pyrrole as a reducing agent has been recently reported [86]: pyrrole oxidation was obtained by mixing graphene oxide with pyrrole at 95°C for 12 h.

Yield <sup>=</sup> SP in ( HSAG SP adduct )

**Table 2.** Preparation of HSAG‐SP adducts: yield of functionalizationa

Variation coefficient, calculated for a population of 10 samples, was 1.7%.

Evaluated from TGA, with the equation reported in the text.

184 Graphene Materials - Structure, Properties and Modifications

details, see Section 8).

a

b

HSAG‐SP‐M (b) and HSAG‐SP‐T (c) are shown in **Figure 2**.

substantially altered by the reaction with SP, even via ball milling.

after the extraction \_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_ SP in (HSAG SP adduct ) before the reaction

.

Very high yield of functionalization (larger than 90%) was achieved, both through mechanical and thermal treatments. In particular, thermal treatment with 0.1 as the SP:HSAG molar ratio gave a yield of 99%. This indicates that SP is able to establish strong interactions with HSAG. It is worth noting that the yield of functionalization decreased with the starting SP:HSAG molar ratio. WAXD and Raman analyses were used to study the organization at the solid state of HSAG‐SP adducts (details are in Ref. [71]). X‐ray diffraction patterns and Raman spectra of HSAG (a),

**Scheme 3.** Block diagram for the preparation of HSAG‐SP adducts, by using either mechanical or thermal energy (for

**Method SP/HSAG molar ratio Yield (%)** Mechanical 0.100 97.3 Thermal 0.100 91.2b Thermal 0.050 79.9 Thermal 0.010 63.6 Thermal 0.005 59.5

As regards WAXD patterns, in HSAG‐SP‐M and HSAG‐SP‐T, (002) reflection remains at the same 2θ value as in pristine HSAG, indicating that SP was not intercalated in the graphitic interlayer space. The number of layers stacked in a crystalline domain, calculated by applying the Scherrer equation to (002) reflection, as explained above, was found to decrease from HSAG (35) to HSAG‐SP‐T (29) to HSAG‐SP‐M (24). Peak shape analysis was as well performed on (100) and (110) reflections, analysing pristine HSAG and the adducts: the correlation length was found in a range from 26.5 to 28 nm. Hence, the in‐plane order of HSAG was not

× 100 (1)

It is worth observing that the relative intensity of the peak at 802 cm−1 decreases in the spec‐ tra of the adducts, in particular in the spectrum of HSAG‐SP‐M adduct. This peak is due to the C=C stretching of the pyrrole ring and such a reduction might lead to hypothesize that C=C bonds are involved in a chemical reaction. Working hypothesis for such a process could be the Diels‐Alder reaction with the graphitic substrate. It is known that graphite is able to react with dienes and dienophiles through Diels‐Alder reaction [27]. The presence of electron

**Figure 3.** FTIR spectra of HSAG (a), HSAG‐SP‐T adduct (b), HSAG‐SP‐M adduct, (c) and SP (d).

withdrawing groups in alpha position with respect to the C=C bond of the pyrrole ring, as it could occur if the methyl group was oxidized to aldehyde, would favour the cycloaddition reaction.

The mechanical adduct HSAG‐SP‐M, with 13% by mass of SP (determined from TGA), was characterized by means of high‐resolution transmission electron microscopy (HRTEM). As it has been reported [71], stable water suspensions (1 mg/mL) were prepared and samples were isolated from supernatant suspensions, after centrifugation for 10 min at 2000 rpm and for 5 and 60 min at 9000 rpm. HRTEM micrographs at lower and higher magnifications are shown in **Figure 4**.

Micrograph in **Figure 4a** reveals that the lateral size of HSAG‐SP‐M adduct is of few hundreds nanometers, that means of the same order of magnitude of pristine HSAG. Two comments can be made: the milling step does not appreciably affect the structure of the graphitic lay‐ ers and SP promotes the dispersion in water of graphene layers with pretty large lateral size. Stacks of graphene layers (indicated in the boxes), disposed perpendicularly to the beam, are in **Figure 4b** and **c**. These figures show the most abundant stacks in populations that contained little larger of lower number of layers. In **Figure 4b**, stacks isolated after centrifuga‐ tion for 10 min at 2000 rpm are of about 3.5–4.2 nm: they are made by about 10–12 stacked graphene layers. Stack in **Figure 4c**, isolated after centrifugation for 5 min at 9000 rpm, is of about 2.8 nm and is made by about eight graphene layers.

Stacks of graphene layers can be fractionated by means of centrifugation; such fractionation appears to be prevailingly due to the number of stacks, rather than to their lateral size. Thanks to the functionalization with SP, nano‐stacks of HSAG can be isolated, with a number of gra‐ phene layers which depends on the adopted experimental conditions.

#### **4.3. Adducts of SP with other graphitic substrates**

Other synthetic graphites were used for preparing adducts with SP. Their main characteristics are shown in **Table 4** in Section 8. Adducts were prepared by giving thermal energy to the graphite / SP mixture and functionalization yield was estimated by means of TGA. Values are shown in **Table 3**.

These findings indicate that the formation of graphite‐SP adducts is not due to the seren‐ dipitous combination of SP with HSAG, but to the affinity and/or reactivity of the pyrrole compound with the graphitic substrate. More in particular, it is worth commenting that, in spite of the different surface area, by using 0.1 as the molar ratio between SP and the graphitic substrate, essentially the same functionalization yield was obtained.

**Figure 4.** Micrographs of HSAG‐SP‐M adduct isolated from supernatant solutions, after centrifugation for 10 min at 2000 rpm (a, b) and 5 min at 9000 rpm (c). Micrographs are low magnification bright field TEM, target: 50 nm (a) and HRTEM image, target: 5 nm (b, c).


**Table 3.** Graphite‐SP thermal adducts: functionalization yield.

withdrawing groups in alpha position with respect to the C=C bond of the pyrrole ring, as it could occur if the methyl group was oxidized to aldehyde, would favour the cycloaddition

**Figure 3.** FTIR spectra of HSAG (a), HSAG‐SP‐T adduct (b), HSAG‐SP‐M adduct, (c) and SP (d).

186 Graphene Materials - Structure, Properties and Modifications

The mechanical adduct HSAG‐SP‐M, with 13% by mass of SP (determined from TGA), was characterized by means of high‐resolution transmission electron microscopy (HRTEM). As it has been reported [71], stable water suspensions (1 mg/mL) were prepared and samples were isolated from supernatant suspensions, after centrifugation for 10 min at 2000 rpm and for 5 and 60 min at 9000 rpm. HRTEM micrographs at lower and higher magnifications are shown

Micrograph in **Figure 4a** reveals that the lateral size of HSAG‐SP‐M adduct is of few hundreds nanometers, that means of the same order of magnitude of pristine HSAG. Two comments can be made: the milling step does not appreciably affect the structure of the graphitic lay‐ ers and SP promotes the dispersion in water of graphene layers with pretty large lateral size. Stacks of graphene layers (indicated in the boxes), disposed perpendicularly to the beam, are in **Figure 4b** and **c**. These figures show the most abundant stacks in populations that

reaction.

in **Figure 4**.


**Table 4.** Synthetic graphites used for the preparation of adducts with serinol pyrrol**e**.

#### **4.4. Adducts of SP with HSAG: what are they?**

Results discussed in previous paragraphs allow commenting as follows. Serinol pyrrole and a graphitic substrate can establish a strong interaction and form very stable adducts. The functionalization of the graphitic substrate with SP leaves the bulk structure of graphene layers substantially unaltered.

The origin of such strength and stability could be attributed to the π‐π interaction of the aromatic moieties, the pyrrole ring in SP and the C6 rings in the graphene layers. However, experimental indications suggest the possibility that SP is able to give rise to a chemical reaction with the aromatic layers of the substrate, more exactly to a cycloaddition reaction. More experimental data would be required to support this hypothesis. At present, rather than stretching too far intriguing inferences, it is worth examining applications that HSAG‐SP adducts allow to pursue.

### **5. Adducts of SP and SP‐based polymers with CB and CNT**

This chapter is focused on the functionalization of graphene layers with a serinol derivative containing a pyrrole ring. In Section 4.3, it has been reported that the formation of stable graphite‐SP adducts does not occur only by using a high‐surface area nanosized graphite, such as HSAG:SP, but is indeed able to interact/react with different types of graphites.

Objective of the present paragraph is to demonstrate that SP is able to interact also with sp<sup>2</sup> carbon allotropes other than graphites. Adducts were formed by SP with carbon allotropes such as carbon black (CB N326) [87, 88] and multiwalled carbon nanotubes. Moreover, adducts with CNT were formed by polyurethane [69, 89] and polyethers [71, 89] oligomers containing SP as a comonomer.

CB‐SP adduct [88] was prepared by using thermal energy, in the absence of solvents or catalysts, as described in Section 8. IR spectrum of the adduct, upon extraction, revealed the presence of peaks characteristic of SP. Water suspensions of CB‐SP adducts were prepared and were first sonicated for 10 min and then centrifuged at 2000 rpm for 5 min. UV‐Vis measure‐ ments were taken after each step of the procedure and related spectra are shown in **Figure 5a**.

UV‐Vis absorption remained substantially the same after centrifugation. Stability of CB‐SP and CB water suspensions were compared after centrifugation: suspensions are in the vials shown in **Figure 5b**. These results indicate that the treatment with SP confers a hydrophilic nature to CB, such as to allow the obtainment of stable water suspensions.

Polyurethanes (PU) oligomers (with molar mass up to about 11 × 10<sup>3</sup> g/mol) were prepared through the solvent‐free polymerization of serinol pyrrole and 1,6‐hexamethylene diisocy‐ anate [69]. Adducts of such oligomers with multiwalled CNT were prepared. Suspensions in acetone of CNT‐PU adducts (prepared by sonication, with 1 mg/mL as the concentration and 46% by mass of PU in the adduct) were stable even after centrifugation. HRTEM analy‐ sis of CNT‐PU adduct revealed prevailingly disentangled CNTs, with intact skeleton, deco‐ rated by PU oligomers, tightly adhered to CNT surface. Micrograph of a CNT‐PU adduct is shown in **Figure 6a**. Extraction tests were performed at room temperature, with ethyl acetate as the solvent, on adducts of CNT with PU oligomers with or without SP (in the latter case, 2,2‐dimethyl‐1,3‐propanediol was the comonomer): mass loss was about 1 and 23% for the respective adducts. It is well known [90] that polyurethanes are very effective CNT modifiers, because they are able to establish very stable interaction, thanks to the π‐π interaction between the carbonyl groups and the aromatic rings. It is thus worth mentioning that SP is able to enhance such interaction, revealing a sort of synergistic effect between different π systems.

**4.4. Adducts of SP with HSAG: what are they?**

layers substantially unaltered.

3807*<sup>a</sup>* Surface‐enhanced

Supplier: Asbury Carbon.

Supplier: Imerys.

a

b

flake graphite

188 Graphene Materials - Structure, Properties and Modifications

adducts allow to pursue.

containing SP as a comonomer.

Results discussed in previous paragraphs allow commenting as follows. Serinol pyrrole and a graphitic substrate can establish a strong interaction and form very stable adducts. The functionalization of the graphitic substrate with SP leaves the bulk structure of graphene

98.9 n. a. 0.36 17.2 2.26

**/g) Density (g/cm3**

**)**

The origin of such strength and stability could be attributed to the π‐π interaction of the aromatic moieties, the pyrrole ring in SP and the C6 rings in the graphene layers. However, experimental indications suggest the possibility that SP is able to give rise to a chemical reaction with the aromatic layers of the substrate, more exactly to a cycloaddition reaction. More experimental data would be required to support this hypothesis. At present, rather than stretching too far intriguing inferences, it is worth examining applications that HSAG‐SP

This chapter is focused on the functionalization of graphene layers with a serinol derivative containing a pyrrole ring. In Section 4.3, it has been reported that the formation of stable graphite‐SP adducts does not occur only by using a high‐surface area nanosized graphite, such as HSAG:SP, but is indeed able to interact/react with different types of graphites.

Objective of the present paragraph is to demonstrate that SP is able to interact also with sp<sup>2</sup> carbon allotropes other than graphites. Adducts were formed by SP with carbon allotropes such as carbon black (CB N326) [87, 88] and multiwalled carbon nanotubes. Moreover, adducts with CNT were formed by polyurethane [69, 89] and polyethers [71, 89] oligomers

CB‐SP adduct [88] was prepared by using thermal energy, in the absence of solvents or catalysts, as described in Section 8. IR spectrum of the adduct, upon extraction, revealed the presence of peaks characteristic of SP. Water suspensions of CB‐SP adducts were prepared and were first sonicated for 10 min and then centrifuged at 2000 rpm for 5 min. UV‐Vis measure‐ ments were taken after each step of the procedure and related spectra are shown in **Figure 5a**.

**5. Adducts of SP and SP‐based polymers with CB and CNT**

**Code Grade C (%) Ash (%) Moisture (%) Surface area (m2**

8427*<sup>a</sup>* HSAG 99.8 n. a. n. a. 330.3 n. a. Nano 24*<sup>a</sup>* HSAG 99.7 0.33 1.93 353.2 n. a.

Timrex SFG6*<sup>b</sup>* Expanded graphite 99.0 0.07 0.10 17.0 0.07

**Table 4.** Synthetic graphites used for the preparation of adducts with serinol pyrrol**e**.

Polyether (PE) oligomers (molar mass up to about 2600 g/mol) were prepared from the reaction of serinol pyrrole with 1,6‐dibromo‐hexane and their adducts with CNT were prepared by sonication (in acetone) [71]. Suspensions (1 mg/mL) were stable over months, even after centrifugation. HRTEM analysis revealed prevailingly disentangled CNTs, with intact skeleton, with PE oligomers wrapped on the CNT surface. Micrograph of a CNT‐PE adduct is shown in **Figure 6b**. Extraction tests at room temperature with ethyl acetate were performed on CNT adducts with PE oligomers with or without SP as comonomer. Mass loss was thus about 24% for the adducts with SP based PE and 98% for the ones with a pluronic surfactant without SP.

**Figure 5.** (a) UV‐Vis spectra of water suspensions of CB‐SP adduct: after sonication (1) and after centrifugation (2) and (b) water suspensions of CB‐SP (1) and of CB (2), after centrifugation.

**Figure 6.** HRTEM micrographs of CNT‐PU adduct (a), PE‐CNT adduct (b). PU and PE contain serinol pyrrole as comonomer.
