Controlling Competitive Radical Pathways: Insights From Aryl Diazonium Electrografting

ABSTRACT

Understanding and controlling the reactivity of radical species remains a major challenge in many chemical reactions, with the goal of steering and precisely tuning key processes. Recognized as a simple and efficient method for the preparation of 2D nanomaterials, diazonium electrografting is a prime example of a functionalization technique that still requires deeper mastery to meet the demands of future bottom‐up approaches. In this work, supported by multiphysics simulations, we achieve an unprecedented understanding of the grafting mechanism involved in nitrobenzene diazonium electrografting and its impact on the resulting nanofilm composition. It is demonstrated that film growth arises from a competition between aryl and diazenyl radicals, leading to the incorporation of variable amounts of azo‐bridged nitrophenyl units. By implementing control strategies using radical scavengers and redox inhibitors, corroborated by simulations, we show that diazenyl radicals are preferentially grafted at the substrate/film interface rather than within the film structure itself. Finally, we demonstrate that selective trapping of aryl radicals over diazenyl radicals enables the formation of azo‐enriched films, thus opening the way for precise tuning of film composition and selective radical reactivity.

This work shows that nitrobenzene diazonium electrografting results from competition between aryl and diazenyl radicals, which govern film composition. Using simulations and control strategies, grafting conditions are tuned to favor specific radicals. It enables selective formation of azo‐enriched films and paves the way toward the precise control of nanofilm structure.

1. Introduction

Radical species are central to a wide array of chemical processes, including material sciences, electrochemistry [1, 2, 3, 4], photochemistry [5, 6], and biology [7, 8, 9]. Their high reactivity makes them both powerful and challenging to control, as they typically exhibit short lifetimes, and are often formed within complex reaction networks involving multiple competing intermediates. Despite significant progress in developing methods to detect and trap radical species, achieving selective control over their reactivity remains a fundamental challenge, especially in systems where several reactive intermediates coexist. Functionalization with aryldiazonium salts exemplifies such complex radical‐based mechanisms. Upon electrochemical reduction, these salts generate highly reactive aryl radicals capable of covalently grafting to surfaces, thereby forming robust organic films [10]. This versatile and widely employed strategy has found applications in areas ranging from carbon fiber modification [11] and nanomedicine [12] to organic electronics [13], biosensing [14], and industrial coatings [15]. However, precise control over the structure, thickness, and composition of the resulting films remains elusive and is closely linked to the underlying radical processes. Several strategies have been proposed to modulate aryldiazonium salt reactivity, including the use of sterically hindered substituents [16], protecting groups [17], and redox cross‐inhibition techniques that suppress radical formation near the surface [18]. Yet, a deeper level of control necessitates a clear understanding of the types of radicals involved and their distinct reactivity profiles. The canonical mechanism involves a concerted electron transfer to the diazonium ion (ArN₂⁺), producing an aryl radical (Ar•), which subsequently reacts with either the substrate or pre‐grafted species to form polyphenylene multilayers [19, 20]. However, the consistent detection of azo linkages (–N = N–) in surface‐bound films‐confirmed by X‐ray photoelectron spectroscopy (XPS) [21], time‐of‐flight secondary ion mass spectrometry (TOF‐SIMS) [22], and Raman spectroscopy [23] indicates the existence of an alternative or complementary reaction pathway [24]. A widely accepted hypothesis suggests that surface‐bound aryl radicals can couple with diazonium cations to form internal Ar–N = N–Ar motifs. [22] Nevertheless, this model does not account for azo bridges observed at the film–substrate interface [25, 26]. Recent studies have presented compelling evidence for the formation of diazenyl radicals (Ar–N = N•) as primary intermediates. These species are generated through a stepwise mechanism in which electron transfer precedes bond cleavage during the reduction of diazonium salts [27] and related compounds such as azosulfates [28]. This insight implies that aryl and diazenyl radicals may coexist and compete during film growth.The present study aims to enhance control over the composition of aryldiazonium‐derived films and to elucidate the mechanisms underpinning their formation. We investigate the influence of key experimental parameters, including diazonium concentration, reduction potential, redox cross‐inhibition, and radical scavengers, on the surface concentration and spatial distribution of azo linkages. Additionally, we employ multiphysics simulations to analyze the reaction pathways associated with each condition, providing a comprehensive understanding of radical selectivity and reactivity and the measurement of rate constants, including the reaction of radicals on carbon surfaces.

2. Results and Discussion

4‐nitrobenzenediazonium (4‐NBD) was chosen as a model diazonium salt in the present work, as its grafting behavior has intensively been characterized in many fundamental studies [21, 29, 30, 31, 32]. We have first examined the films obtained by electrochemical reduction of 4‐NBD on glassy carbon via XPS measurements and cyclic voltammetry to confirm the presence of reduced nitrogen after the grafting procedure under our modification conditions, as illustrated in Figure 1 (experimental information for chemicals, electrochemistry, and XPS is detailed in section S1). Chronoamperometric modifications of a glassy carbon electrode were carried out at ‐0.4 V vs Ag/AgNO₃ in a 0.1 M nBu₄NPF₆ CH₃CN solution containing 1 mM of 4‐NBD. This potential corresponds to the minimum driving force that permits reaching the maximum and steady state surface concentration of the nitrophenyl functions (vide infra, Figure 3a). The cyclic voltammogram recorded in KOH 0.1 M after the grafting presents a first reduction peak arising at −960 mV vs Ag/AgNO₃ during the forward scan, followed by a less intense oxidation peak centered at −410 mV vs Ag/AgNO₃ during the backward scan. This electrochemical behavior is characteristic of the nitrophenyl group grafted on the electrode [33], the cathodic peak being ascribed to both the irreversible reduction (6 e⁻/6 H⁺) of the nitro group into amino groups and to the partial reduction into the hydroxylamine intermediate as seen in Figure 1. The latter can be subsequently reversibly oxidized into the nitroso derivative through a reversible 2 e⁻/2 H⁺ process (anodic peak at –480 mV vs Ag/AgNO₃). Therefore, considering that all NO₂ moieties are electrochemically active, one can extract the apparent surface concentration Γ of the film grafted on the surface from both the reduction and the oxidation contributions, using equation 1:

$$\mathrm{\Gamma }=\frac{-{\mathrm{Q}}_{\mathrm{red}}+{\mathrm{Q}}_{\mathrm{ox}}}{\mathrm{nFA}}$$ (1)

where Q_red and Q_ox being the charge of the reduction and oxidation peaks, respectively, n = 6 the total number of electrons passing through the reduction and the oxidation processes, F the Faraday constant and A the geometric area of the electrochemical surface. In the case of a nitrophenyl coverage, Γ = 6–7 × 10⁻¹⁰ mol.cm⁻² corresponds to the grafting of one monolayer of nitrophenyl moieties [34].

FIGURE 1.

Illustration of the electroreduction of the 4‐nitrobenzenediazonium leading to the competitive grafting of the diazenyl/aryl radicals on carbon substrate, and strategy based on the combination of CV and XPS measurements to extract the proportion of azo bonding within the grafted film. The CV was recorded in KOH 0.1 M at 50 mV/s on a 3 mm diameter glassy carbon electrode after a grafting for 5 min at ‐0.4 V performed in a CH₃CN solution containing 1 mM of 4‐NBD and 0.1 M nBu₄NPF₆. Corresponding XPS core level spectrum was recorded on a 0.5 cm² glassy carbon plate modified in the same conditions.

FIGURE 3.

(a) Evolution of the nitrophenyl surface concentration for films prepared on GC electrodes in a 1 mM solution of 4‐NBD in CH₃CN 0.1 M nBu₄NPF₆ for 5 min as a function of the potential fixed for the grafting (vs Ag/AgNO₃), and sigmoidal fitting of the values ranging from‐0.5 to ‐1.8 V below. (b) Normalized XPS N1s core‐level spectra of modified GC at ① (+0.1 V) and ② (‐1.6 V). (c) Illustration of the preferential grafting of the diazenyl radical species and reduction of the aryl radical at a potential below ‐0.8 V (vs Ag/AgNO₃).

Besides, X‐ray photoelectron spectroscopy was exploited to determine the composition of the nitrophenyl films. The N1s core level spectra of the modified surfaces exhibit two principal peaks located at 406 and 400 eV (Figure 1, see also Figure S1 for detailed wide and core level spectra). The peak observed at 406 eV can unambiguously be attributed to the nitro groups [35] whereas the peak at 400 eV is assigned to the presence of azo bridges within the layer [22]. Additionally, a peak at ∼403 eV is observed and can be attributed to protonated amines [36, 37].

The presence of ‐N = N‐ links was also further supported by spectroscopic techniques, including UV‐vis, Raman, and infraRed measurements as shown in section S3. Noteworthy, their combination confirms the presence of azo bridges within the film, whereas SERS measurements suggest the formation of azo bridges at the surface/film interface.

The combination of electrochemical and XPS measurements was employed to determine the impact of the diazonium concentration on the final surface concentration of azo bridges within the nanofilm. A series of chronoamperometric modifications at ‐0.4 V vs Ag/AgNO₃ was carried out using various concentrations of 4‐NBD, ranging from 10⁻⁶ to 10⁻² M. The XPS results obtained on a glassy carbon electrode after electrochemical grafting of the 4‐NBD for 5 min, as well as their electrochemical behavior recorded in KOH 0.1 M are shown in Figures 2a and 2b respectively. Upon increase of the diazonium concentration, an increase of both the contributions at ∼ 406 eV and ∼ 400 eV corresponding to the NO₂ and –N = N‐ bonds is observed, whereas the one at ∼403 eV attributed to the protonated amines remains mainly constant. This trend is accompanied by an increase of the electrochemical current corresponding to the NO₂ reduction at ‐1 V. More quantitatively, Figure 2c plots the evolution of the percentage of Ar‐NO₂ groups attached via N = N bonding extracted from the XPS spectra (Figure 2a) as a function of the nitrophenyl surface concentration extracted from the CV as described above. From a value of 4×10⁻¹⁰ mol.cm⁻² recorded for 10⁻⁶ M, the nitrophenyl surface concentration increases up to 25×10⁻¹⁰ mol.cm⁻² for 10⁻³ M, a value for which it seems to reach a steady state value. This expected trend reflects an increasing aryl radical production at the electrode interface and thus an increasingly efficient grafting. For the higher concentrations, reaching a maximum value illustrates the self‐limited grafting process due to the formation of the passivating organic film [33, 38].

FIGURE 2.

(a) XPS N1s core‐level spectra of modified GC with increasing 4‐NBD concentration. (b) Corresponding first cycles of the CVs recorded at 100 mV/s (potential vs Ag/AgNO₃) in KOH 0.1 M of the Ar‐NO₂ reduction and Ar‐NHOH/Ar‐NO oxidation process from which Γ is extracted. (c) Scheme of the 4‐NBD electroreduction considering a stepwise mechanism. (d) Evolution of Γ on previously modified GC electrodes (in red) and their corresponding estimated azo bond percentages (in black) at various diazonium concentrations and simulated trends, respectively in light red and grey, using k_g,diazenyl = 4.105 M⁻¹.s⁻¹ and k_g,aryl = 1.103 M⁻¹.s⁻¹. (e) Electrochemical quartz microbalance experiment and corresponding evolution of Γ modeled for k_g,diazenyl = 4.105 M⁻¹.s⁻¹ and k_g,aryl = 1.104 (green), 1.103 (yellow), 1.102 (blue), and 10 (purple).

Interestingly, an increase of the percentage of nitrophenyl groups attached via azo bridges, from 20 ± 3% to 43 ± 3%, was observed when the solution was diluted 1000‐fold (i.e., from 10⁻³ M to 10⁻⁶ M). In the frame of the previously proposed mechanism to explain the azo bridge formation [22], for which diazonium attacks of radical intermediates formed on the top of the film are involved, increasing the diazonium concentration should result in a higher, or at least constant, percentage of azo bridges within the nanofilm. The opposite trend highlighted by our results suggests that the formation of azo bridges is rather due to the direct grafting of a nondenitrogenated species as the diazenyl radical associated to a stepwise electroreduction of 4‐NBD recently evidenced [27], contrary to what was previously published [39]. Briefly, in this mechanism schematized in Figure 2c, the 4‐NDB is reduced at the electrode surface through a one‐electron reduction process leading to the formation of a diazenyl radical as a first intermediate. The diazenyl radical can then either be grafted onto the electrode or decomposed spontaneously to an aryl radical that can be grafted. This sequence of reactions had already been proposed during the pulse radiolysis study of the reduction of diazonium salts by solvated electrons in water [40]. Diazenyl radicals have also been observed as intermediates during the photolysis of arylazosulfones [41], during the paired electrosynthesis of azo compounds [42] and the diazenylation of active methylene compounds [43].

To further validate this hypothesis and deeply understand the trends of the diazonium concentration on the layer composition, numerical simulations were performed based on the mechanism shown in Figure 2d. In this model, detailed in section S4, the grafting process of both the diazenyl radical and the aryl radical was considered analogous to adsorption processes inspired by previous work [44]. The competition between the grafting of the diazenyl and the aryl radicals was investigated by studying the effect of the grafting rate constants k_g,diazenyl, k_g,aryl on both the total surface coverage and the percentage of azo bonds within the film at various diazonium concentrations. Considering this mechanism, it is possible to reproduce the experimental trends on both the surface coverage and the percentage of azo bonds even if a multitude of couples of constants for 1.10⁵ M⁻¹.s⁻¹< k_g,diazenyl < 1.10⁶ M⁻¹.s⁻¹ and k_g,aryl aryl > 10 M⁻¹.s⁻¹ could reasonably fit the experimental results. Examples of couples of constants are shown in Figure S5.

To refine those constants, we further investigated their effect on the overall grafting dynamic and compared them to EQCM experiments that probed the growth of the layer during the grafting process presented in Figure 2e (experimental details for EQCM are presented in Section S1). As shown in Figure S6, k_g,diazenyl has a limited effect on the film growth dynamic during the first 10 s and will only influence the final surface concentration that slightly decreases if k_g,diazenyl < 100 M⁻¹.s⁻¹. Contrariwise, k_g,aryl largely influences this dynamic that fits with the EQCM experiments for k_g,aryl ∼1.10³ M⁻¹.s⁻¹ as shown in Figure 2e. Based on these findings, the effect of diazonium concentration can be accurately reproduced using the grafting rate constants k_g,diazenyl = 4.10⁵ M⁻¹.s⁻¹ and k_g,aryl = 1.10³ M⁻¹.s⁻¹. Taking these constants into account, an initial grafting rate of ∼1.5 nmol·cm⁻²·s⁻¹ was obtained for the diazenyl species, a value similar to that previously reported for aryl grafting under similar experimental conditions [45], whereas a much lower rate (0.03 nmol·cm⁻²·s⁻¹) is obtained for the grafting of the aryl species. These results highlight that the variation in the N = N ratio arises from competition between the grafting of the diazenyl radical and that of the aryl radical, the latter being formed via decomposition of the former. It can be further explained by comparing the simulated growth of diazenyl and aryl films at both high and low concentrations. Figure S7, which plots Γ_diazenyl and Γ_aryl (respectively, the surface concentration of nitrophenyl grafted via azo bridging and directly grafted) over time for 1 mM and 0.001 mM diazonium using the rate constants determined above, shows that decreasing the diazonium concentration leads to a less efficient grafting of aryl groups (∼80% decrease from 1 to 0.001 mM) compared to diazenyl groups (∼65% decrease). Therefore, this favors the grafting of diazenyl radicals over aryl radicals when the diazonium concentration decreases.

We further investigated the impact of the electrochemical driving force applied during grafting on the structure of the formed film, particularly considering the previously demonstrated reduction of radical species, especially the aryl radical, into their anionic counterparts [46]. Therefore, a series of chronoamperometric modifications in a 1 mM 4‐NBD solution was conducted within a potential ranging from 0.4 to ‐1.6 V vs Ag/AgNO₃. Figure 3a shows the evolution of Γ_tot as a function of the grafting potential. In the +0.4 V to ‐0.4 V range, an increasing trend can be observed up to a value of 25.10⁻¹⁰ mol.cm⁻². This expected behavior reflects the increasing production of radicals and consequently the growth of the layer thickness. The stabilization of the layer expansion observed around ‐0.5 V is characteristic of the self‐limited grafting process caused by the insulating property of the organic layer [21]. Then, a drop of the Γ_tot is visible from ‐0.8 V to ‐1.8 V vs Ag/AgNO₃, attesting to a change in the electrografting regime. This drop aligns with a > 10 times increase of the average reduction current (i_av) recorded during the electrografting between ‐0.6 V and –1.8 V vs Ag/AgNO₃ (chronoamperograms in Figure S8). This increase cannot be solely attributed to the oxygen reduction reaction (as shown in Figure S8), the contribution of which is negligible under our experimental conditions. The additional reduction current, already observed previously with other aryl diazoniums [47], most likely results from the electroreduction of radicals (i.e., diazenyl and/or aryl) into their corresponding anions at the electrode interface. This may also explain the observed decrease in film thickness. Additional XPS analyses were performed after the grafting at +0.1 and ‐1.6 V vs Ag/AgNO₃). As shown in Figure S9, the N1s core level spectra revealed that the percentage of N = N within the film slightly decreases from +0.1 to ‐0.4 V vs Ag/AgNO₃ from 22 ± 3% to 19 ± 3% and then increases up to 38 ± 3% at ‐1.6 V vs Ag/AgNO₃. Additional simulations were performed by considering either the aryl radical reduction or the diazenyl reduction based on the scheme in Figure 2c as detailed in section S6. In both cases, the experimental trend for Γ_tot and the azo bond percentage versus the potential can be qualitatively reproduced, as shown in Figure S9. Moreover, the apparent half‐wave potential E_{1/2, app} ∼ 1.2 V ± 0.2 V vs Ag/AgNO₃ extracted from both, the current increase and the drop of Γ, aligns well with the reduction potentials of aryl radicals, previously estimated to be E° = ‐1.01 and ‐1.08 V/ Ag/AgNO₃ for phenyl and 4‐nitrophenyl radicals (see Figure 3a) [47].

As explained in section S4, the model considers a maximum surface concentration of the diazenyl Γ_{diazenyl,max lim}  =  5×10⁻¹⁰ mol.cm⁻², based on the total surface concentration and the percentage of N = N obtained experimentally for high diazonium concentrations (1 to 10 mM). This boundary condition, necessary to explain the experimental trends of Figure 2d, reinforces the hypothesis that the diazenyl is rather confined at the electrode surface, in line with the presence of N = N at the material surface observed by Raman and IR. It suggests a more efficient grafting of the diazenyl on the electrode surface than on the already grafted film compared to the aryl radical.

To obtain accurate information on the distribution of azo links in the thickness of the film and thus discriminate between the diazenyl and the aryl radical grafting, we exploited a previously described strategy to lower the layer thickness. In this strategy, 2,2‐diphenylpicrilhydrazyl (DPPH) can be exploited as a redox inhibitor that is reduced at the electrode and reacts with aryl diazonium cations in solution [18, 48]. This cross‐reaction, schematized in Figure 4a, results in the delocalization of the radical production at a given distance from the electrode surface, which prevents their coupling with the electrode. Initially employed to precisely control the thickness of the grafted molecules [18], this approach was more recently used to bring some quantitative insights on the diazonium reduction mechanism [49].

FIGURE 4.

(a) Mechanism involved in the redox‐cross inhibition of the grafting using DPPH as a redox inhibitor. (b) Evolution of the nitrophenyl surface concentration (in red) and percentage of azo bridges (in black) for films prepared on GC electrodes in a 1 mM solution of 4‐NBD in CH₃CN 0.1 M nBu₄NPF₆ for 5 min as a function of DPPH concentration.

A series of chronoamperometric modifications at ‐0.4 V vs Ag/AgNO₃ in solutions containing 1 mM of 4‐NBD and various concentrations of DPPH was achieved. Figure 4b shows the evolution of the nitrophenyl surface concentration in parallel with the corresponding percentages of azo bridges as a function of DPPH concentration. Both XPS analysis and cyclic voltammetry are shown in Figure S10. Without DPPH, a maximum surface coverage of 25×10⁻¹⁰ mol.cm⁻² was reached, corresponding to a multilayered grafting of 2 ± 0.5 nm (i.e. 3–4 layers considering a size of 0.8 nm for a nitrophenyl moiety) in which 20% of the nitrophenyl are attached via azo bridges (Figure S11, see also section S1 for atomic force microscopy experimental details) [38]. For increasing DPPH concentrations, a drop of the surface coverage can be observed, accompanied by a marked increase of the azo links proportion in the layer in qualitative agreement with the results obtained by decreasing the diazonium concentration (Figure 2). However, the use of 2 equivalents of DPPH, which allows to strictly stop the grafting at the monolayer stage [50], led to the attachment of 82 ± 3% of the nitrophenyl groups via azo bridges. This result clearly demonstrates that the diazenyl radical preferentially reacts with the carbon surface rather than with attached phenyl rings, as suggested in the previous section.

To understand the mechanism behind those trends, numerical simulations were attempted (details in section S8). First, a simple model involving the sole redox cross inhibition within the stepwise mechanism was built following the reaction scheme in Figure S12. By fixing the parameters of the stepwise mechanism as defined previously, the effect of the rate constant k_f of the reaction between the reduced redox inhibitor and the aryl diazonium was investigated.

It can be observed in Figure S12 that an increase of k_f leads to a limited decrease of the nitrophenyl surface concentration (Γ_tot) when the DPPH concentration increases compared to the experimental one, even for very high values of k_f (up to 10⁷ M⁻¹.s⁻¹). This slight decrease can be explained by the fact that only Γ_diazenyl decreases, whereas Γ_aryl remains constant, as shown in Figure S12. This also leads to a decrease of the N = N percentage, which is the opposite of the experimental trend. Such inadequacy demonstrates that this simple mechanism is not sufficient to explain our results.

In a recent study, we suggested that the redox cross‐reaction can be accompanied by a trapping of the radical by the redox molecule acting as a radical scavenger [49]. Therefore, as schematized in Figure 4a, we further considered the reaction between the DPPH and the aryl/diazenyl radical in the simulation. As a first approximation, the same rate constant of DPPH with both radicals was considered. As shown in Figure S13, one can nicely reproduce the general trends using k_f  =  10⁶‐10⁷ M⁻¹.s⁻¹ and k_p  =  10⁶‐10⁷ M⁻¹ (k_f = 10⁶ M⁻¹.s⁻¹ and k_p = 10⁶–10⁷ M⁻¹.s⁻¹ in Figure 4b), where k_p is the reaction rate constant for the coupling between the DPPH and the diazenyl/aryl radicals. This strongly suggests that DPPH, in its radical form, can efficiently interact with aryl/diazenyl radicals during its diffusion toward the electrode prior to reduction. It is, however, interesting to note that for C_DPPH  =  0.5 mM, Γ obtained experimentally is significantly lower than the simulated one. It is due to the fact that the DPPH concentration should normally not be sufficient to limit the growth of the film within this range of time (300s). One possible explanation involves the self‐degradation of aryl and/or diazenyl radicals, either through dimerization or via radical attack on phenyl groups in the bulk. This can occur either by reaction with diazonium salts through a Gomberg–Bachmann‐like mechanism [51] or through dimerization/polymerization processes involving the radicals themselves. For instance, Figure S14 shows that, when considering the dimerization of aryl radicals as an example, the experimental results are well reproduced. A similar trend is obtained when considering the Gomberg–Bachmann‐like mechanism.

Finally, since it is suggested that the radical form of the DPPH can react with aryl/diazenyl radicals, we further investigated the radical scavenging strategy to drive the structure of the film and modulate the ratio of azo bridges by selectively trapping one of the two radicals during the deposition process. To this end, glassy carbon electrodes were first modified by chronoamperometry at +0.1 V in the presence and in the absence of DPPH. At this potential, DPPH is under its radical form (see Figure S15 for the voltametric study), which allows the trapping of free radicals in solution [52]. It was observed that the surface generated with and without DPPH have similar nitrophenyl surface concentrations, approximately 10×10⁻¹⁰ mol.cm⁻². However, following the addition of the radical scavenger in the grafting solution, the percentage of nitrophenyl groups attached via azo bridges determined by XPS (see spectra in Figure S16) increases from 27 ± 3% to 51 ± 3% (Figure 5). This result strongly suggests that the DPPH preferentially traps the aryl radical over the diazenyl radical.

FIGURE 5.

(a) Histogram shows the effect of the radical scavengers on the percentages of nitrophenyl groups anchored via azo bridges for films prepared on GC electrodes in a 1 mM solution of 4‐NBD in CH₃CN 0.1 M nBu₄NPF₆ for 5 min without (black) and with (grey) the addition of 1 mM DPPH or PBN. (b) Stepwise mechanism involving the trapping of the radicals by the radical scavengers. (c) Simulated evolution of the ratio of the azo bonds concentration with (CN = N) and without (CN = N,0) the radical scavenger versus (k_p), the reaction rate constant associated with the trapping of the aryl/diazenyl radicals by the radical scavengers. Experimental values obtained for the PBN and DPPH are added based on the results shown in a).

To confirm the effectiveness of this approach, a second modification was carried out at high driving force (i.e. ‐1.6 V) in a condition allowing to generate a similar nitrophenyl surface concentration (i.e., 13 ×10⁻¹⁰ mol.cm⁻²). In this experiment, N‐terbutyl‐α‐phenylnitrone (PBN), which is not reduced at low potential (see Figure S17), was used as a radical scavenger [53] instead of DPPH. Following a similar trend, the percentage of azo bridges significantly increased from 38 ± 3% without PBN to 66 ± 3% in the presence of PBN, as shown in the histogram of Figure 5 (see also XPS spectra in Figure S16). These consistent results confirm that both DPPH and PBN trap preferentially aryl radicals, leading to the formation of films enriched in azo bridges. Notably, such a trapping strategy is expected not to be limited to these two radical scavengers; other compounds, particularly nitrone derivatives, could also be considered.This radical scavenging process was further simulated using the mechanism displayed in Figure 5b, as detailed in Section S10. As shown in Figure 5c, based on the ratio of the percentage of ‐N = N‐ obtained with and without the radical scavengers, the reaction rate between the DPPH and the aryl/diazenyl was found to be one order of magnitude higher than that involving PBN (k_p,DPPH ∼ 10⁶ M⁻¹.s⁻¹ and k_p,PBN ∼ 10⁵ M⁻¹.s⁻¹) considering that the radical trapping is a diffusion‐limited process. In that configuration, the increase of ‐N = N‐ percentage observed in the presence of the radical scavengers can be explained by a preferential reaction of the radical scavengers toward the aryl radical. Indeed, as shown in Figure S17, even for high values of k_p,x (up to 10⁹ M⁻¹.s⁻¹, with x being either DPPH or PBN), Γ_diazenyl remains unchanged, whereas Γ_aryl becomes negligible. As shown in Figure S18, this can be explained by the high grafting rate of the diazenyl compared to the aryl since a decrease of k_g,diazenyl would lead to a decrease of Γ_diazenyl.

3. Conclusion

In summary, our results shed new light on the electrografting mechanism of 4‐nitrobenzenediazonium salts onto carbon substrates. By combining electrochemical characterization, XPS and spectroscopic analysis, EQCM, and numerical simulations, we provide compelling evidence for the coexistence and competitive grafting of both diazenyl and aryl radicals. Contrary to the commonly accepted denitrogenation pathway, our findings highlight the significant contribution of nondenitrogenated diazenyl radicals, particularly at low diazonium concentrations or under conditions limiting radical diffusion (e.g., in the presence of DPPH). The modeling accurately reproduces the experimental trends by considering distinct grafting rates for both radical species and suggests that azo link formation predominantly occurs through direct grafting of diazenyl species at the electrode surface. These insights revise the current understanding of multilayer formation during diazonium electrografting and provide a predictive framework to modulate film composition and structure by tuning experimental parameters such as concentration, potential, or the use of redox inhibitors. Work is underway to validate and extend the methodology presented in this study to other diazonium salts. More broadly, this work proposes strategies to achieve radical selectivity that could be of interest in many chemical and biochemical systems where competing radical species may interfere. To further advance the understanding of the role of the substrate, a valuable next step would be to identify the impact of the carbon surface structure on the electroreduction mechanism of diazonium salts. It is indeed likely, as previously reported [54], that the stabilization of azo species at the electrode surface depends on the degree of graphitic character of the carbon surface.

Conflicts of Interest

The authors declare no conflict of interest

Supporting information

The authors have cited additional references within the Supporting Information [27, 55, 56, 57, 58, 59, 60, 61, 62, 63].

Helis S., Pinson J., Decorse P., Hamon J., Noël J.‐M., and Breton T., “Controlling Competitive Radical Pathways: Insights From Aryl Diazonium Electrografting.” Chemistry – A European Journal 32, no. 5 (2026): e03001. 10.1002/chem.202503001