Improving Kinetics of “Click-Crosslinking” for Self-Healing Nanocomposites by Graphene-Supported Cu-Nanoparticles

Investigation of the curing kinetics of crosslinking reactions and the development of optimized catalyst systems is of importance for the preparation of self-healing nanocomposites, able to significantly extend their service lifetimes. Here we study different modified low molecular weight multivalent azides for a capsule-based self-healing approach, where self-healing is mediated by graphene-supported copper-nanoparticles, able to trigger “click”-based crosslinking of trivalent azides and alkynes. When monitoring the reaction kinetics of the curing reaction via reactive dynamic scanning calorimetry (DSC), it was found that the “click-crosslinking” reactivity decreased with increasing chain length of the according azide. Additionally, we could show a remarkable “click” reactivity already at 0 °C, highlighting the potential of click-based self-healing approaches. Furthermore, we varied the reaction temperature during the preparation of our tailor-made graphene-based copper(I) catalyst to further optimize its catalytic activity. With the most active catalyst prepared at 700 °C and the optimized set-up of reactants on hand, we prepared capsule-based self-healing epoxy nanocomposites.

In the field of ROMP-based self-healing [7,9], the curing behavior of renewable norbornenyl-functionalized isosorbide monomers in the copolymerization with dicyclopentadiene (DCPD) was investigated exhibiting a higher reactivity, consequently facilitating low temperature ROMP. Additionally, higher crosslinking densities were observed, resulting in improved thermal and mechanical properties highlighting the potential of renewable ROMP-monomers towards self-healing applications [9]. software (V 3.40, Anton Paar Germany GmbH, Ostfildern, Germany, 2008) and OriginPro 8G (Version 8.0951, OriginLab Corporation, Northampton, MA, USA, 2008) were used.
Freeze drying: Freeze drying was performed on a LyoQuest freeze dryer from Telstar (Utrecht, The Netherlands) operating at −80 • C and 0.18 mbar.
Ultrasonicator: For the dispersion of the GO-species via ultrasonication a sonication tip Vibra Cell VCX500 from Zinsser Analytic (Frankfurt, Germany) was used.
Flame atomic absorption spectroscopy (FAAS): FAAS was performed on a novAA 350 #113A0641 Tech: Flamme spectrometer from Analytik Jena AG (Jena, Germany) using Aspect LS 1.4.1.0 (Analytik Jena AG, Jena, Germany) as software. Therefore, external calibration and calibration via doping were performed. To determine the copper-content within TRGO-Cu 2 O the samples were burned to ash at 800 • C under atmospheric conditions and were dispersed in nitric acid (2 M). This solution was diluted in a 1:1 ratio with a potassium chloride solution (0.2%) and a mixture containing 25 mL of the dispersed nitric sample solution and 25 mL of the potassium chloride solution.
Transmission electron microscopy (TEM): TEM investigations were performed using a EM 900 transmission electron microscope from Carl Zeiss Microscopy GmbH (Oberkochen, Germany) and the images were taken with a SSCCD SM-1k-120 camera from TRS (Moorenweis, Germany). For sample preparation, TRGO-Cu 2 O was dispersed in water and sprayed on a carbon-layered copper grid. After one minute, the excess solution was removed with filter paper and the samples were dried at room temperature.
X-ray diffraction (XRD): XRD-measurements were performed on a D8 X-ray diffractometer from Bruker AXS GmbH (Karlsruhe, Germany). For analysis of the raw data, Diffrac. Suite EVA 3.1 (Bruker AXS GmbH, Karlsruhe, Germany) with an integrated database for the determination of the phases was used as software. For sample preparation, TRGO-Cu 2 O was rubbed in the presence of isopropanol and was put on a glass slide. For evaluation of data, OriginPro 8G (Version 8.0951, OriginLab Corporation, Northampton, MA, USA, 2008) was used.

General Synthesis Procedure for the Preparation of Trivalent Azides
The synthesis was carried out under a dry atmosphere of nitrogen. A two-necked round-bottom flask equipped with magnetic stir bar, rubber septum and gas tap was heated under vacuum and flushed with nitrogen several times. 4-Dimethylaminopyridine (0.2 eq) was added to 2 (1.0 eq) dissolved in dry DMF and the solution was stirred for ten minutes at room temperature. Afterwards, the desired anhydride or acid chloride (6.0-8.0 eq) was added dropwise to the reaction mixture and the solution was stirred at room temperature. After finishing the reaction, the crude product was either purified by extraction or column chromatography and the obtained product was dried in high vacuo. As determined via NMR-spectroscopy, final products contain up to 25% impurities such as free epichlorohydrine and bivalent residues as trimethylolpropane triglycidyl ether was used in technical grade (see 1 H-NMR spectrum in Supplementary Materials Figure S1).

Synthesis of Trivalent Alkyne and Trivalent Azides
In Figure 1, an overview of the synthesized trivalent alkyne 1 and the trivalent azides is given. Trivalent alkyne 1 and trivalent azide 2 have been synthesized according to literature [22,58,59].

Synthesis of Trivalent Alkyne and Trivalent Azides
In Figure 1, an overview of the synthesized trivalent alkyne 1 and the trivalent azides is given.
Trivalent alkyne 1 and trivalent azide 2 have been synthesized according to literature [22,58,59]. The synthesis of the trivalent azides 3*, 4, 4*, 5 and 5* is presented in Scheme 1, and details are given in the general synthesis procedure in Section 2.3 and in the Supplementary Materials. In brief, the trivalent azide 2 was either converted with acetic anhydride or the desired acid chloride at room temperature in dry DMF and in the presence of DMAP as a nucleophilic basic catalyst. The synthesis of the trivalent azides 3*, 4, 4*, 5 and 5* is presented in Scheme 1, and details are given in the general synthesis procedure in Section 2.3 and in the Supplementary Materials. In brief, the trivalent azide 2 was either converted with acetic anhydride or the desired acid chloride at room temperature in dry DMF and in the presence of DMAP as a nucleophilic basic catalyst.

Synthesis of Trivalent Alkyne and Trivalent Azides
In Figure 1, an overview of the synthesized trivalent alkyne 1 and the trivalent azides is given.
Trivalent alkyne 1 and trivalent azide 2 have been synthesized according to literature [22,58,59]. The synthesis of the trivalent azides 3*, 4, 4*, 5 and 5* is presented in Scheme 1, and details are given in the general synthesis procedure in Section 2.3 and in the Supplementary Materials. In brief, the trivalent azide 2 was either converted with acetic anhydride or the desired acid chloride at room temperature in dry DMF and in the presence of DMAP as a nucleophilic basic catalyst. The trivalent azide 3* was obtained as a light-yellow, viscous liquid (85% azide content) containing the bivalent azide and free epichlorohydrine present in the starting material (trimethylolpropane triglycidyl ether (technical grade)). This mixture was further studied without purification to investigate its suitability for easy preparable and up-scalable room temperature-based self-healing nanocomposites relying on "click-crosslinking" reactions. For the synthesis of the azides 4 and 5, a pure trivalent compound (4 and 5) and a mixture of bi-and trivalent product (4* and 5*, 85 and 75% azide content, respectively) were obtained and further separated via column chromatography. All prepared multivalent azide-and alkyne-functionalized compounds were characterized via NMRand IR-spectroscopy as well as ESI-TOF mass spectrometry proving their purity and functional group content (for more details see Supplementary Materials Figures S2-S5: 1 H-and 13 C-NMR spectra of trivalent azides 4 and 5).

DSC Investigation of "Click-Crosslinking" Trivalent Alkyne and Trivalent Azides
Thermal analysis by differential scanning calorimetry (DSC) provides useful information about the relationship between the extent of a (crosslinking) reaction and the required time of curing at a certain temperature. Furthermore, information about kinetic parameters can be retained. Thus, DSC analysis is helpful to obtain a wide range of data of the investigated "click-crosslinking" reactions of trivalent alkyne 1 and trivalent azides 3*, 4, 4*, 5 and 5* such as the enthalpy of the reaction (∆H), the onset temperature (T onset ), the temperature at the maximum of the DSC curve (T p ) and the apparent activation energy of the reaction (E a, app ). It should be mentioned, that the experimentally determined activation energies are indicated as apparent activation energies, since it is well known that the physical conditions during crosslinking are at least partially restricted by the mass and heat transport in the solid phase, consequently influencing the internal energy as well as the vibrational states of the investigated reactants [60,61]. Furthermore, it should be emphasized that our prepared TRGO-Cu 2 O catalyst is composed of several graphene sheets showing a lack in dispersibility with the reactants due to missing functional groups of increasing polarity. Thus, a relatively high apparent activation energy is expected-a phenomenon also observed in graphene oxide nanocomposite epoxy coatings [62].
Via DSC investigations, the "click-crosslinking" reaction conversion can be estimated with respect to a determined reference value, a maximum ∆H value (262 kJ·mol −1 ) for a reference click reaction between phenylacetylene and benzyl azide when quantitatively forming one triazole unit and therefore being representative of one successful "click" reaction. This reference value is in line with reported literature values for "click" reactions ranging between 210 to 270 kJ·mol −1 [63]. DSC measurements for the "click-crosslinking" reaction of trivalent alkyne 1 with trivalent azide 3* investigated at a heating rate of 5 K·min −1 in the presence of different homo-and heterogenous copper(I) catalysts (1 mol % per functional group) as well as without catalyst (W/O) are plotted in Figure 2 and the obtained results are summarized in Table 1.     Table 1. Thermal properties, reaction temperatures (T onset and T p ), reaction enthalpies (∆H), apparent activation energies (E a, app ) and conversions of the "click-crosslinking" reaction of trivalent alkyne 1 and trivalent azide 3*, 4, 4*, 5 or 5* with different catalysts (1 mol %) as well as without catalyst (W/O) at a heating rate of 5 K·min −1 . TRGO-Cu 2 O was prepared at 600 • C.  1 According to our previous publication [22] and our experience, the error is typically ≈ ±5 K. 2 According to our previous publication [22], the error is typically ≈ ±6 kJ·mol −1 . 3 Calculated with respect to the enthalpy for 100% click conversion which is ∆H = 262 kJ·mol −1 for the reference reaction of phenylacetylene and benzyl azide with 1 mol % of Cu(PPh 3 ) 3 Br as catalyst.
For the Huisgen cycloaddition of trivalent alkyne 1 and trivalent azide 3* (Table 1, Entry 1) 78% conversion was achieved corresponding to an enthalpy of 205 kJ·mol −1 . Crosslinking took place at high temperatures and T onset and T p were observed at 91 and 133 • C, respectively. In the case of the homogenous catalysts (Cu(PPh 3 ) 3 F and Cu(PPh 3 ) 3 Br, Table 1, Entry 2 and 3), the observed enthalpies were 191 and 185 kJ·mol −1 , consequently showing a conversion of 73 and 71%, respectively. By using TRGO-Cu 2 O as a catalyst for "click-crosslinking" 1 and 3* (Table 1, Entry 4), an enthalpy of 177 kJ·mol −1 was observed corresponding to 67% conversion. Moreover, the lowest apparent activation energy (55 kJ·mol −1 ) and the lowest maximum peak temperature T p (63 • C) were achieved in the presence of TRGO-Cu 2 O and the lowest T onset (39 • C) by using Cu(PPh 3 ) 3 F. According to these results, TRGO-Cu 2 O and Cu(PPh 3 ) 3 F were the best catalysts for the "click-crosslinking" reaction of trivalent alkyne 1 and trivalent azide 3*.
To investigate the activity of the pure trivalent azides 4 and 5 in comparison to the partially functionalized trivalent azides 4* and 5* (85 and 75% azide content) together with trivalent alkyne 1 in the CuAAC crosslinking reaction, DSC measurements were run by using TRGO-Cu 2 O as a catalyst (1 mol % per functional group). The DSC thermograms at 5 K·min −1 with TRGO-Cu 2 O as a catalyst are plotted in Figure 3a,b and the obtained results are summarized in Table 1.
For "click-crosslinking" trivalent alkyne 1 with trivalent azide 4 ( Table 1, Entry 5), the observed enthalpy was 233 kJ·mol −1 , corresponding to 89% conversion. In comparison, the observed enthalpy for "click-crosslinking" trivalent alkyne 1 with trivalent azide 4* (Table 1, Entry 9) showed a slightly lower enthalpy value of 211 kJ·mol −1 and a conversion of 81%, related to the presence of the bivalent byproduct consequently lowering the conversion. In contrast, the reaction temperatures (T p and T onset ) were decreased for "click-crosslinking" trivalent alkyne 1 with trivalent azide 5*, mainly attributed to a lower viscosity and therefore, to a faster diffusion. Thus, an enthalpy of 248 kJ·mol −1 was observed for "click-crosslinking" trivalent alkyne 1 with trivalent azide 5 (Table 1, Entry 10) at relatively high reaction temperatures corresponding to 95% conversion. In comparison, a lower enthalpy of 127 kJ·mol −1 (48% conversion) was measured for the "click-crosslinking" reaction of trivalent alkyne 1 with trivalent azide 5* (Table 1, Entry 14) while the reaction temperatures were reduced (T onset = 32 • C, T p = 56 • C). Consequently, further investigations towards easily up-scalable room temperature-based self-healing nanocomposites were continued by using trivalent azides 4* and 5*. To investigate the activity of the pure trivalent azides 4 and 5 in comparison to the partially functionalized trivalent azides 4* and 5* (85 and 75% azide content) together with trivalent alkyne 1 in the CuAAC crosslinking reaction, DSC measurements were run by using TRGO-Cu2O as a catalyst (1 mol % per functional group). The DSC thermograms at 5 K•min −1 with TRGO-Cu2O as a catalyst are plotted in Figure 3a,b and the obtained results are summarized in Table 1. For "click-crosslinking" trivalent alkyne 1 with trivalent azide 4 ( Table 1, Entry 5), the observed enthalpy was 233 kJ•mol −1 , corresponding to 89% conversion. In comparison, the observed enthalpy for "click-crosslinking" trivalent alkyne 1 with trivalent azide 4* (Table 1, Entry 9) showed a slightly lower enthalpy value of 211 kJ•mol −1 and a conversion of 81%, related to the presence of the bivalent byproduct consequently lowering the conversion. In contrast, the reaction temperatures (Tp and Tonset) were decreased for "click-crosslinking" trivalent alkyne 1 with trivalent azide 5*, mainly attributed to a lower viscosity and therefore, to a faster diffusion. Thus, an enthalpy of 248 kJ•mol −1 was observed for "click-crosslinking" trivalent alkyne 1 with trivalent azide 5 (Table 1, Entry 10) at relatively high reaction temperatures corresponding to 95% conversion. In comparison, a lower enthalpy of 127 kJ•mol −1 (48% conversion) was measured for the "click-crosslinking" reaction of trivalent alkyne 1 with trivalent azide 5* (Table 1, Entry 14) while the reaction temperatures were reduced (Tonset = 32 °C, Tp = 56 °C). Consequently, further investigations towards easily up-scalable room temperaturebased self-healing nanocomposites were continued by using trivalent azides 4* and 5*. DSC thermograms of "click-crosslinking" trivalent alkyne 1 with trivalent azides 4* or 5* without catalyst (Huisgen cycloaddition) as well as with different catalysts (1 mol % per functional group) are illustrated in Figure 3c,d, respectively. For the uncatalyzed crosslinking reaction of trivalent alkyne 1 with trivalent azide 4* (Table 1, Entry 6), a high reaction enthalpy of 227 kJ·mol −1 and high reaction temperatures (T onset = 94 • C, T p = 125 • C) were observed, corresponding to a conversion of 87%. For the homogenous catalyst Cu(PPh 3 ) 3 F the lowest enthalpy was observed, while the click reaction happened at relatively low temperatures (  Entry 7 and 8), respectively. When reacting trivalent alkyne 1 with trivalent azide 5*, lower conversions were observed for the (uncatalyzed) Huisgen cycloaddition reaction as well as for all "click-crosslinking" reactions. Thus, a conversion of 66% was achieved for the uncatalyzed reaction (Table 1, Entry 11), corresponding to an enthalpy of 174 kJ·mol −1 , while "click-crosslinking" happened at high temperatures (T onset and T p of 102 and 141 • C). In comparison to the other catalysts, TRGO-Cu 2 O (Table 1, Entry 14) resulted in the lowest enthalpy of 127 kJ·mol −1 , which was achieved at T onset and T p of 32 and 56 • C, respectively. In the case of the homogeneous catalysts (Cu(PPh 3 ) 3 F and Cu(PPh 3 ) 3 Br) a similar conversion of 53 to 57% was observed, while lower reaction temperatures were achieved in the presence of Cu(PPh 3 ) 3 F (Table 1, Entry 13, T onset = 38 • C, T p = 62 • C). According to the obtained results, Cu(PPh 3 ) 3 F turned out to be the best catalyst for the "click-crosslinking" reaction of trivalent alkyne 1 and trivalent azides 4* or 5*.
3.3. DSC Investigation of "Click-Crosslinking" Trivalent Alkyne 1 and Trivalent Azides 3*, 4* and 5* at 0 • C We were interested in quantifying "click-crosslinking" reactions during preparation and mixing of the components, thus understanding whether "click" reactions at 0 • C play an essential role. Therefore, DSC measurements were applied to investigate the kinetic behavior of the "click-crosslinking" reaction between trivalent alkyne 1 and trivalent azides 3*, 4* and 5* with different chain lengths to find a suitable and fast catalytic system for the CuAAC. As the usage of Cu(PPh 3 ) 3 F resulted in high enthalpies at low crosslinking temperatures in the previously performed experiments, Cu(PPh 3 ) 3 F was chosen as catalyst to check the activity of the multivalent azides and alkynes within "click-crosslinking" at 0 • C. Therefore, 1:1 mixtures of trivalent alkyne 1 and different trivalent azides (3*, 4* or 5*) together with Cu(PPh 3 ) 3 F (1 mol % per functional group) were prepared. Immediately after preparation of the mixtures, DSC measurements were run at a heating rate of 5 K·min −1 to observe the enthalpy of the "click-crosslinking" reaction at time zero (∆H 0 ). Afterwards, the mixtures were stored at 0 • C and further DSC investigations were conducted in defined time intervals and the conversion was determined (for more details see Supplementary Materials Figure S4 and Table S1).
Immediately after mixing trivalent alkyne 1 and trivalent azide 3*, the measured enthalpy of the "click-crosslinking" reaction was 186 kJ·mol −1 , and T onset and T p were 39 and 67 • C, respectively ( Figure 4, black squares; Supplementary Materials, Table S1, Entry 1). After 48 h storage at 0 • C, the "click-crosslinking" enthalpy decreased around one half of its initial value and reached 82 kJ·mol −1 , corresponding to 56% conversion (Supplementary Materials, Table S1, Entry 3). After 312 h, the "click-crosslinking" enthalpy decreased further to 24 kJ·mol −1 (13 days, 87% conversion, Supplementary Materials, Table S1, Entry 9). Afterwards, the conversion did not show any significant increase, and the "click-crosslinking" reaction between alkyne 1 and azide 3* reached its maximum conversion of 93% at 0 • C after 576 h (24 days, Supplementary Materials, Table S1, Entry 12).     The mixture of trivalent alkyne 1 and trivalent azide 4* showed a high enthalpy of 222 kJ·mol −1 immediately after sample preparation, and T onset and T p were 37 and 54 • C, respectively (Figure 4, red circles; Supplementary Materials, Table S1, Entry 14). After 48 h, the "click-crosslinking" enthalpy decreased to 80 kJ·mol −1 and a conversion of 64% was observed (Supplementary Materials, Table S1, Entry 15). After 384 h, the conversion of the "click-crosslinking" reaction at 0 • C increased further to 85%, corresponding to a reaction enthalpy of 33 kJ·mol −1 (Supplementary Materials, Table S1, Entry 18). In comparison, the conversion after 624 h did not significantly change and reached a constant value of 88% (27 kJ·mol −1 , Supplementary Materials, Table S1, Entry 19). Thus, it was concluded that the maximum "click-crosslinking" conversion obtainable by converting trivalent alkyne 1 and trivalent azide 4* at 0 • C is below 90%.
Immediately after preparation of the reaction mixture of trivalent alkyne 1 and trivalent azide 5*, a "click-crosslinking" enthalpy of 172 kJ·mol −1 was observed, and T onset and T p were 63 and 86 • C, respectively (Figure 4, blue curve; Supplementary Materials, Table S1, Entry 20). After 96 h, the reaction enthalpy of the "click-crosslinking" reaction decreased to 89 kJ·mol −1 , related to a conversion of 48% (Supplementary Materials, Table S1, Entry 23). After 552 h, the enthalpy decreased further to 8 kJ·mol −1 (95% conversion), and T onset was not detectable anymore due to the very low reaction enthalpy. After 648 h, a complete "click-crosslinking" conversion was achieved, and no reaction peak was observed.
Moreover, in all crosslinking experiments investigated at 0 • C, the peak temperature T p and the onset temperature T onset increased in comparison to their initial values. This increase of the crosslinking temperatures is mainly attributed to the early network formation taking place at 0 • C, and is therefore related to a slowed down monomer diffusion.
To sum up the DSC investigations at 0 • C, the "click-crosslinking" reactions of trivalent alkyne 1 with trivalent azides 3* and 4* at 0 • C were faster within the first 100 h than the corresponding "click-crosslinking" reaction in the presence of trivalent azide 5*. This phenomenon was mainly attributed to the increasing chain length of the trivalent azide. Thus, under the same conditions, molecules with shorter side chains show faster "click-crosslinking" in comparison to the molecules with the longer chain length, mainly attributed to lower viscosity. Nevertheless, slightly higher conversions of the "click-crosslinking" reaction were observed on long timescales for the trivalent azides with increased chain length.

Rheology Investigation of "Click-Crosslinking" Trivalent Alkyne 1 and Trivalent Azides 4* and 5*
The viscoelastic and the kinetic behavior of the "click-crosslinking" reaction between trivalent alkyne 1 and trivalent azides 4* and 5* (1:1 ratio of azide and alkyne) and the resulting self-healing capability were investigated via in situ rheology. Therefore, the isothermal "click-crosslinking" reaction was directly performed on a rheometer plate at 20 • C using Cu(PPh 3 ) 3 F (1 mol % per functional group) as a catalyst. The observed crossover times for the "click-crosslinking" reaction of trivalent alkyne 1 and trivalent azides 4* and 5* with Cu(PPh 3 ) 3 F were 190 and 1445 minutes, respectively (see Supplementary Materials, Figure S6). By comparing these times with the crossover time of the "click-crosslinking" reaction of trivalent azide 3* with the trivalent alkyne 1 which was 35 minutes [22], it was concluded that with increasing chain length of the azide the crossover time increased. This observation was in line with the DSC investigations proving a decreased "click-crosslinking" reactivity with increasing chain length.

Synthesis and Characerization of TRGO-Cu 2 O Prepared at Different Temperatures
TRGO-Cu 2 O was prepared via thermal reduction of copper(II)-modified graphene oxide in a glass tube furnace according to a previously published procedure [21,64,65], while the reduction temperature was varied between 300 to 800 • C, finally obtaining six different batches of the desired heterogeneous copper(I) catalyst (for more details see Scheme S6 in the Supplementary Materials) to optimize the synthesis procedure in terms of the catalytic activity since DSC. Investigations for "click-crosslinking" trivalent alkyne 1 with trivalent azide 3*, 4* or 5* revealed slightly higher reaction temperatures (T onset and T p ) in comparison to the homogeneously catalyzed "click-crosslinking" reactions in the presence of Cu(PPh 3 ) 3 F.
The prepared TRGO-based copper(I) catalysts were investigated via XRD-measurements (see Supplementary Materials Figure S7a) and for all prepared TRGO-Cu 2 O catalysts the characteristic reflex of GO at 2θ = 11 • has disappeared due to successful reduction. Furthermore, for all prepared catalysts, reflexes at 2θ = 38 • and 2θ = 43 • were observed related to the formed copper species (pure copper as well as copper(I)). For the TRGO-Cu 2 O samples prepared at 700 and 800 • C, two additional reflexes at 2θ = 51 • and 2θ = 26 • were detected as characteristic reflexes for pure copper and graphite, respectively. Thus, it could be concluded, that some of the oxidic groups have been eliminated during thermal reduction partially resulting in graphite-like structures. The broad signal at 2θ = 25 • observed for all prepared TRGO-based copper(I) catalysts was caused by an interference with the sample holder.
The prepared TRGO-Cu 2 O samples were further analyzed via TEM investigations (see Supplementary Materials Figure S7b-g) in which the formed nanosized copper(I) particles were visualized. The size of the particles was investigated via Image J. The particle size increased with increasing preparation temperature of TRGO-Cu 2 O, and average particle-diameters between 25 to 150 nm were determined. Furthermore, it was observed, that TRGO-based catalysts prepared at 700 and 800 • C displayed a more disperse distribution of nanosized copper(I) particles, which may be attributed to the formation of pure copper interacting with graphite-like structures detected in XRD investigations.

Crosslinking Reactions of Alkynes and Azides in the Presence of TRGO-Cu 2 O Prepared at Different Temperatures
The catalytic activity of the TRGO-Cu 2 O catalysts (prepared at different temperatures) towards "click-crosslinking" was investigated via DSC investigations. Therefore, in the first step, a model reaction between phenylacetylene and benzyl azide (1:1 ratio of azide and alkyne) was investigated at a heating rate of 5 K·min −1 , and the reaction temperatures (T onset and T p ), the reaction enthalpy (∆H) and the conversion were recorded (for more information see Supplementary Materials Figure S8 and Table S2). While comparing the different catalysts prepared at 300 to 800 • C, it was observed that T onset and T p decreased with increasing temperature applied during the reduction of copper(II)-modified graphene oxide towards TRGO-Cu 2 O, in line with the expectation and the increasing size of the formed copper particles. Thereby, one exception was observed, and the catalyst prepared at 500 • C showed the highest peak temperature. Thus, FAAS measurements were performed to determine the loading of TRGO with immobilized copper nanoparticles. While most of the prepared catalysts displayed around 8 wt % of copper, a strong decrease was noted for the TRGO-Cu 2 O synthesized at 500 • C, directly linked to the observed reduced catalytic activity during the click reaction of phenylacetylene and benzyl azide.
In the next step, the catalytic activity was tested in a more complex system suitable for the preparation of room-temperature based self-healing epoxy nanocomposites. "Click-crosslinking" of trivalent alkyne 1 and trivalent azide 3* (1:1 ratio of azide and alkyne, assuming 66% azide content of 3*) in the presence of different TRGO-Cu 2 O catalysts prepared from 300 to 800 • C (5 wt %) at a heating rate of 5 K·min −1 (see Supplementary Materials Figure S10) was investigated, and the obtained reaction temperatures (T onset and T p ), the reaction enthalpies (∆H) and the conversions are summarized in Table 2. For comparison, the non-catalyzed reaction between trivalent alkyne 1 and trivalent azide 3* was repeated, showing similar reaction temperatures (T onset = 96 vs. 91 • C, T p = 130 vs. 133 • C; Table 2, Entry 1) as described before, but a reduced conversion due to the higher amount of bivalent residues (66 vs. 75% azide content, see Supplementary Materials Figure S7). All DSC measurements were performed in three independent experiments to ensure their reproducibility and to approximate the expected error. Thereby, differences especially in the peak shape were observed to be related to sample preparation and limited blending of the reactants as well as a limited diffusion to the catalyst surface being typical for a reaction directly catalyzed by a solid support [60,61]. Table 2. Thermal properties of the "click-crosslinking" reaction of trivalent alkyne 1 and trivalent azide 3* with TRGO-Cu 2 O as a catalyst (prepared at different temperatures) at a heating rate of 5 K·min −1 : Reaction temperatures (T onset and T p ), reaction enthalpies (∆H) and conversions.

Entry
Catalyst For "click-crosslinking" trivalent alkyne 1 and trivalent azide 3*, a similar trend was observed as in the model reaction between phenylacetylene and benzyl azide. Thus, the reaction temperatures (T onset and T p ) decreased with increasing reduction temperature applied during the preparation of the different TRGO-Cu 2 O catalysts. Thereby, the lowest T onset and T p of 64 and 80 • C, respectively, were observed for "click-crosslinking" trivalent alkyne 1 and trivalent azide 3* in the presence of TRGO-Cu 2 O prepared at 700 • C while the catalyst prepared at 500 • C showed the worst result.
The obtained peak temperatures for the click model reaction of phenylacetylene and benzyl azide as well as for the "click-crosslinking" reaction of trivalent alkyne 1 and trivalent azide 3* were correlated to the preparation temperature of TRGO-Cu 2 O and the corresponding amount of copper within these catalysts (see Figure 5). Table 2. Thermal properties of the "click-crosslinking" reaction of trivalent alkyne 1 and trivalent azide 3* with TRGO-Cu2O as a catalyst (prepared at different temperatures) at a heating rate of 5 K•min −1 : Reaction temperatures (Tonset and Tp), reaction enthalpies (ΔH) and conversions.

Entry
Catalyst Tonset 1 (°C) Tp 1 (°C) ∆H 2 (kJ⋅mol −1 ) Conversion 3 (%) For "click-crosslinking" trivalent alkyne 1 and trivalent azide 3*, a similar trend was observed as in the model reaction between phenylacetylene and benzyl azide. Thus, the reaction temperatures (Tonset and Tp) decreased with increasing reduction temperature applied during the preparation of the different TRGO-Cu2O catalysts. Thereby, the lowest Tonset and Tp of 64 and 80 °C, respectively, were observed for "click-crosslinking" trivalent alkyne 1 and trivalent azide 3* in the presence of TRGO-Cu2O prepared at 700 °C while the catalyst prepared at 500 °C showed the worst result.
The obtained peak temperatures for the click model reaction of phenylacetylene and benzyl azide as well as for the "click-crosslinking" reaction of trivalent alkyne 1 and trivalent azide 3* were correlated to the preparation temperature of TRGO-Cu2O and the corresponding amount of copper within these catalysts (see Figure 5). Please note that the lines between the measuring points are drawn to guide the eye.
The already mentioned trends within the catalytic activity of TRGO-Cu2O were well highlighted for both click reactions. The catalyst prepared at 500 °C showed a low amount of copper and The already mentioned trends within the catalytic activity of TRGO-Cu 2 O were well highlighted for both click reactions. The catalyst prepared at 500 • C showed a low amount of copper and consequently a high peak temperature during the click reactions, while the best results in terms of copper loading as well as peak temperature were observed for TRGO-Cu 2 O prepared at 700 • C. The decrease of the copper loading for the catalyst prepared at 500 • C may be related to the decrease of oxygen-functional groups with increasing reduction temperature known to act as reactive sites for the nucleation and growth of metal nanoparticles [66,67]. Following this argumentation, a further decrease in the copper loading together with a decreased catalytic activity would be expected, which was not observed in this particular study. We therefore assume that either the formation of pure copper indicated by XRD investigations is boosting the catalytic activity or that the diffusion of metal atoms, especially favored at higher temperatures, leads to the formation of non-stable metal clusters influencing the catalytic activity [68].
Thus, for the preparation of self-healing nanocomposites according to a previously published procedure [19], the optimized TRGO-Cu 2 O catalyst prepared at 700 • C was used. Thereby, the trivalent alkyne 1 together with this particular TRGO-Cu 2 O were directly distributed within the epoxy matrix together with µm-sized capsules filled with trivalent azide 3*. Further self-healing investigations of our optimized healing system as well as the determination of self-healing efficiencies are ongoing in our laboratories and will be part of a future publication.

Conclusions
Different low molecular weight, multivalent azides with small structural changes were synthesized and their crosslinking kinetics was investigated in a CuAAC-based curing reaction. Therefore, different homogeneous and heterogeneous copper(I) catalysts were screened and the kinetic parameters such as the reaction temperatures, the enthalpy of the reaction as well as the apparent activation energies were recorded via DSC investigations. We observed, that the "click-crosslinking" reactivity decreased with increasing chain length of the azide. Furthermore, a significant click reactivity of all investigated azides could be proven already at 0 • C.
The reaction conditions for the preparation of our home-made TRGO-Cu 2 O catalyst were optimized: When increasing the reaction temperature to 700 • C, the resulting copper(I) catalyst displayed the highest catalytic activity as shown in model click reactions as well as in "click-crosslinking" reactions between trivalent alkyne 1 and trivalent azide 3*.
The tuned catalyst was subsequently dispersed in an epoxy matrix together with the trivalent alkyne 1 and the encapsulated trivalent azide 3* (µm-sized capsules).
Further self-healing investigations of the so prepared capsule-based self-healing graphene-supported epoxy nanocomposites are ongoing and will be part of a future publication.