RESEARCH ARTICLE www.small-methods.com Introducing Design Strategies to Preserve N-Heterocycles Throughout the On-Surface Synthesis of Graphene Nanostructures Maria Tenorio, Cesar Moreno,* Manuel Vilas-Varela, Jesús Castro-Esteban, Pol Febrer, Miguel Pruneda, Diego Peña,* and Aitor Mugarza* the undoped counterparts. It implies, Despite the impressive advances in the synthesis of atomically precise among others, the tunability of the elec- graphene nanostructures witnessed during the last decade, advancing in tronic properties,[1,2] the generation of compositional complexity faces major challenges. The concept of introducing metal-free catalysts,[3–8] the complexation the desired functional groups or dopants in the molecular precursor often fails with transition metal and lanthanides for single atom catalysts[9,10] and pho- due to their lack of stability during the reaction path. Here, a study on the ton emitters,[11] and the affinity enhance- stability of different pyridine and pyrimidine moieties during the on-surface ment of target species for sensing and synthesis of graphene nanoribbons on Au(111) is presented. Combining filtering,[12–15] as well as of alkaline atoms bond-resolved scanning tunneling microscopy with X-ray photoelectron for energy storage applications.[10,16] spectroscopy, the thermal evolution of the nitrogen dopants throughout the Most of these applications require a control on the density, distribution, whole reaction sequence is tracked. A comparative experimental and ab initio and bonding configuration of dopants electronic characterization confirms the presence of dopants in the final that is hardly achievable by top-down structures, revealing also that the pyridinic nitrogen leads to a significant methods. The latter is particularly criti- band downshift. The results demonstrate that, by using synthetic strategies to cal, since nitrogen dopants can acquire lower the reaction temperatures, one can preserve specific N-heterocycles up to eight bonding configurations in throughout all the reaction steps of the synthesis of graphene nanoribbons graphene nanostructures.[17,18] Indeed, the different methods of post-synthesis and beyond the interibbon coupling reaction that leads to addition of nitrogen dopants lead to mul- nanoporous graphene. tiple bonding configurations and their random distribution at irregular edges or pores.[15,19–21] 1. Introduction Bottom-up methods such as the on-surface synthesis (OSS) and the use of specifically designed molecular precursors can of- Adding nitrogen atoms to the edges and pores of graphene-based fer a solution to simultaneously control both the atomic struc- nanostructures is an effective strategy to provide a manifold of ture and the chemical composition.[22] The best example is the interesting properties that amplify the potential applications of bottom-up synthesis of graphene nanoribbons (GNRs), where M. Tenorio[+] , C. Moreno, P. Febrer, M. Pruneda, A. Mugarza C. Moreno Catalan Institute of Nanoscience and Nanotechnology (ICN2) Departamento de Ciencias de la Tierra y Física de la Materia Condensada CSIC and The Barcelona Institute of Science and Technology Universidad de Cantabria Bellaterra, Barcelona 08193, Spain Santander 39005, Spain E-mail: cesar.moreno@unican.es; aitor.mugarza@icn2.cat M. Vilas-Varela, J. Castro-Esteban, D. Peña Centro de Investigación en Química Biolóxica e Materiais Moleculares The ORCID identification number(s) for the author(s) of this article (CiQUS) can be found under https://doi.org/10.1002/smtd.202300768 Departamento de Química Orgánica Universidade de Santiago de Compostela [+] Presentaddress: IMDEA Nanoscience, Cantoblanco, Madrid 28049, Santiago de Compostela 15782, Spain Spain E-mail: diego.pena@usc.es © 2023 The Authors. Small Methods published by Wiley-VCH GmbH. A. Mugarza This is an open access article under the terms of the Creative Commons ICREA Institució Catalana de Recerca i Estudis Avançats Attribution-NonCommercial-NoDerivs License, which permits use and Barcelona 08010, Spain distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. DOI: 10.1002/smtd.202300768 Small Methods 2023, 2300768 2300768 (1 of 10) © 2023 The Authors. Small Methods published by Wiley-VCH GmbH 23669608, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/smtd.202300768 by Readcube (Labtiva Inc.), Wiley Online Library on [16/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.small-methods.com several heteroatoms and functional groups have been introduced before reaching the sublimation temperature, and cannot there- with the same atomic precision as the rest of the structural fore be deposited on the surface. A MALDI characterization of the parameters.[2,23–33] However, these studies also reflect that, in par- material recovered from the evaporator crucible suggests that the ticular when located at the edges, the functional groups in most unwanted polymerization may be driven by the nucleophilic at- cases suffer from isomerization,[33] chemical transformation,[25] tack of the pyridine moieties to the electrophilic 10,10’ positions or partial[30,31] and even total cleavage[34] during the thermally as- of the biantracene core. In contrast, the less basic character of sisted on-surface synthesis. pyrimidines reduces the nucleophilicity of the N lone pairs, pre- Taking into account the aforementioned stability issues, venting the the polymerization of monomers 3 and 4 before subli- and with the aim of extending the reaction sequence to the mation. The stability of the pyrimidine group in these monomers synthesis of supraribbon nanostructures, we have carried out throughout the on-surface synthesis could therefore be studied a systematic study of the stability of N-heterocycles throughout by tracking the thermal evolution of intermediates and final prod- the on-surface synthesis of graphene nanoribbons and their ucts, as shown in the following. coupling into nanoporous graphene (NPG). In particular, we Figure 2 summarizes the complementary STM and XPS anal- have focused on the synthesis of N-doped counterparts of the ysis carried out for each reaction step in order to track the struc- 7–13 AGNR and the corresponding NPG that can be obtained by tural and chemical evolution of monomer 3. After its deposition using a bisphenyl substituted bianthracene precursor.[35,36] We on the Au(111) surface at room temperature, large islands of well- find that pyridine functionalized precursors are less thermally ordered, self-assembled precursor monomers are observed by stable than their pyrimidine counterparts. The reactivity of the STM (Figure 2a), denoting a high diffusivity and substantial inter- former in fact leads to thermally induced polymerization prior molecular interactions of these molecules on Au(111). Close-up to sublimation in ultra-high vacuum (UHV), impeding the images reveal chain like structures made of protrusion pairs that deposition of precursors on the surface. In contrast, the less fit with the staggered conformation of the DBBA core.[36,38] The basic pyrimidine groups are thermally stable enough to permit interlobe lateral distance of 7.4 Å and the apparent height of 2.3 their deposition and withstand all the reaction path up to the Å indicate that the monomer adsorbs with the pyrimidine groups synthesis of GNRs. In order to facilitate cross-coupling reactions at the lower side interacting with the Au surface, as found for the that could lead to supraribbon structures as the NPG,[35] we phenyl group of the undoped DP-DBBA (see Figure S16, Sup- introduce a chlorine atom as substituent in the pyrimidine porting Information).[36] The superimposed model in the inset group, with the idea of replacing the dehydrogenative cross- of Figure 2a suggests that intrachain interactions are mediated coupling reaction by a Cl-based Ullmann coupling that takes by the pyrimidine groups, whereas Br⋅⋅⋅H is the dominant inter- place at lower temperature. The hierarchy between the two chain interaction. Ullmann coupling reactions given by the different C-Br and The integrity of the monomer is confirmed by the XPS re- C-Cl dissociation energies maintain the polymerization and sults of Figure 2d–f (see SI for more quantitative information on cross-coupling reactions in non-overlapping thermal windows. the XPS data analysis). We find that the C 1s core level nicely Our conclusions are founded on a comprehensive scanning fits by imposing the stoichiometric relation of bonding compo- tunneling microscopy (STM), X-ray photoelectron spectroscopy nents, C-C(H):C-NH:C-Br = 28:6:2. For simplicity, C-C(H), C- (XPS) and density functional theory (DFT) study that correlates NH, C-Br refer respectively to C[C2 H]/C[C3 ], C[CNH]/C[N2 H] structural, chemical and electronic information and provides and C[C2 Br], the former two merging a pair of components into atomic scale insight into mechanisms that impede or promote single fitting peaks due to the difficulty to resolve them within the functionalization of graphene nanostructures by OSS. our experimental conditions. The binding energies of the com- ponents obtained from this fit also agree with that of the pris- 2. Results and Discussion tine molecule. C-C(H) and C-Br lie at 284.3 eV and 284.8 eV, re- spectively, very close to values obtained for the undoped molecule Four different N-doped GNR precursors were studied, all adsorbed on Au(111),[36] and the value of 285.4 eV obtained for bianthracene derivatives substituted in 10,10’ positions with C-NH is characteristic of pyridine groups.[39,40] The Br 3d dou- bromine atoms and N-heterocycles in 2,2’ positions (Figure 1), blet appears at 70.7/69.7 eV, energies corresponding to C-bonded together with their undoped counterpart DP-DBBA (10,10’- Br[36,41] For N, we find a single 1s component at 398.8 eV, energy dibromo-2,2’-diphenyl-9,9’-bianthracene) parent.[35–37] In expected for pyridinic N,[17] with relative area close to the stoi- monomers 1 and 2, the two phenyl groups at the 2,2’ posi- chiometric relation. Altogether they represent signatures of the tions of monomer DP-DBBA are replaced by two pyridine rings pristine molecule. (4- and 3-pyridinyl, respectively). In monomer 3, the number of After postannealing to 150°C, monomers polymerize into 1D N-dopants is increased by replacing the two phenyl substituents chains, as can be seen in the STM images of Figure 2b. One in DP-DBBA with two pyrimidines instead (pyrimidin-5-yl). Fi- can attribute the polymerization to the Ullmann coupling of nally, in monomer 4, we introduce two chloropyrimidine groups monomers by using the complementary information obtained in the bianthracene core, in particular two 2-chloropyrimidin-5- from the XPS analysis: the 2.1 eV upshift of the Br 3d doublet yl substituents. is a clear indicative of cleaved, Au-bonded Br (Figure 2e).[36] This The integrity of the functional groups is already challenged is in agreement with the better fit, we obtain when we turn C- at the first stage of the on-surface synthesis procedure, that is Br bonds into C-C by fixing the C 1s multicomponent relation to the sublimation of the precursor monomer in ultra-high vacuum C-C(H):C-NH = 30:6 (Figure 2d). (UHV) for its deposition on the catalytic surface. We find that The morphology and distribution of the polymer chains are pyridine functionalized monomers 1 and 2 already polymerize in stark contrast to that of their undoped counterparts (i.e., Small Methods 2023, 2300768 2300768 (2 of 10) © 2023 The Authors. Small Methods published by Wiley-VCH GmbH 23669608, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/smtd.202300768 by Readcube (Labtiva Inc.), Wiley Online Library on [16/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.small-methods.com Figure 1. Schematic representation of DP-DBBA, studied in Refs. [35–37], dipyridine derivatives 1 and 2, and dipyrimidine derivatives 3 and 4. The final structures achieved with DP-DBBA in a previous work,[35] and with monomer 4 in this work, are sketched at the bottom. In red, bonds formed by Br-based Ullmann couplings; in blue, bonds formed by cyclodehydrogenations; in green, interibbon Cl-based Ullmann couplings. DP-DBBA). The pinning of the latter at the elbow disloca- pyrimidine group (directly or indirectly) promotes the cyclodehy- tions of the herringbone leads to chains with an average length drogenative cyclization of the polymer chain. Although this reac- of 70 nm aligned to the zigzag direction of the herringbone tion cannot be directly tracked in the evolution of the XPS spec- reconstruction.[37] Conversely, the N-doped polymers appear tra due to the proximity of the C-H and C-C components of the mostly confined within the fcc/hcp tracks of the reconstruction, C 1s core level, the presence of the residual H can be inferred with an average length of 18 ± 8 nm (with a total sample of 78 from the absence of the Br core level, which in many other stud- polymers measured), very close to the characteristic length of the ies has been attributed to the desorption of HBr.[36,42,43] The sur- fcc/hcp segments of the zigzag pattern. These differences sug- vival of the pyrimidine groups to this reaction step is supported gest that N-doped polymers interact more strongly with the Au by the fact that the N 1s peak remains at the same energy and surface, and as consequence their length is limited rather than relative area (Figure 2f), although the twofold increase in width guided by the reconstruction pattern, similar to that found for already indicates the onset of some transformations, as it will be other GNRs.[31] discussed later. After further postannealing to 300°C polymer chains, charac- The presence of the N-dopants at the cata-fused pheripheral terized by staggered protrusions in the STM images, appear now rings as well as their site distribution can be explored in detail as planar structures, with an apparent height of 2.2 Å and the by a comparative analysis of bond-resolved (BR)-STM images. triple bay edge structure that characterize the 7–13 AGNRs (see The example displayed in Figure 3a shows clear contrast differ- Figure 2c).[35] The formation of GNRs at 100°C lower tempera- ences between the edge rings of undoped and N-doped GNRs. ture as compared to the undoped counterpart denotes that the Those of the N-doped GNRs appear with lower contrast, as has Small Methods 2023, 2300768 2300768 (3 of 10) © 2023 The Authors. Small Methods published by Wiley-VCH GmbH 23669608, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/smtd.202300768 by Readcube (Labtiva Inc.), Wiley Online Library on [16/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.small-methods.com Figure 2. On-surface synthesis of edge nitrogen-doped graphene nanoribbons from monomer 3, and characterization of electronic properties. a–c) STM images (overview and representative closeview) of the stepwise reaction to grow edge nitrogen-doped graphene nanoribbons. From left to right: as-deposited monomers, 1D polymers, and graphene nanoribbons. d–f) XPS analysis of the C 1s (d), Br 3d (e), and N 1s (f) core levels for monomer 3 at different temperatures of postannealing. STM parameters are listed in Table S8 (Supporting Information). been observed for other N-heterocycle containing structures.[26,45] Similar to the case of polymers, the presence of N-dopants at A more detailed anaylsis of the appearance of edge rings in the the edges seems to determine the GNR distribution on the sur- N-doped GNRs reveals asymmetric features that are inverted at face. Instead of evenly dispersed by the interaction with the re- opposing edges. The asymmetry mimics that imprinted by the constructed Au surface as observed for the undoped GNRs,[37] nitrogen heteroatoms to a fused pyrimidine ring, which is a ro- N-doped GNRs appear bundled, suggesting dominating attrac- bust evidence of their integrity. Their homogeneous distribution tive interibbon interactions. By superimposing models to an in- along each edge of the ribbon, on the other hand, is a direct proof teracting pair of GNRs, as done in Figure 3b, we conclude that of the formation of homochiral GNRs, which is a crucial ingre- the attractive interaction is driven by an effective N ⋅⋅⋅ H pairing dient for the interibbon interaction and the potential formation between pyrimidines (4 N⋅⋅⋅H pairs per pyrimidine). Such N⋅⋅⋅H of nanoporous graphene by their lateral coupling, as pictured in interaction configuration can only be found between pyrimidines Figure 3b. of homochiral pairs of GNRs, which in turn is the configuration Small Methods 2023, 2300768 2300768 (4 of 10) © 2023 The Authors. Small Methods published by Wiley-VCH GmbH 23669608, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/smtd.202300768 by Readcube (Labtiva Inc.), Wiley Online Library on [16/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.small-methods.com Figure 3. Identification of nitrogen atoms in pyrimidine functional groups. a) BR-STM image of a single N-doped GNR, with a zoom in one the outer pyrimidine rings, and the undoped phenyl counterpart. Schemes of the outer aromatic rings have been overlapped for the identification of nitrogen atoms. b) BR-STM image of a couple of interacting N-doped GNRs, with structural models that would correspond to homochiral and heterochiral pairs, showing that only the former would lead to stable bonding configuration. STM parameters are listed in Table S8 (Supporting Information). that would lead to the energetically favorable covalent bonding proposed degradation path). This degradation resembles that pre- in the formation of NPG. By tracing back chirality at each step viously described for tetrathienoanthracene derivatives.[48] The of the formation of GNRs, we note that the homochiral distribu- gradual degradation of pyrimidine groups is in agreement with tions of interacting GNRs are enabled by the fact that enantiopure the XPS data, where the main variation of the N 1s core level peak interactions determine the assembly of both monomer and poly- in the 300–350°C range is an increase in the width that could ac- mer intermediates. The homochiral polymer ensemble shown in count for the different N environments, and at 400°C the peak Figure 2b is one example of it, as can be inferred from the same undergoes a significant reduction of around 30–40% of its area staggering sequence derived from the lobes. (Figure 5b). The charge accumulation induced by the higher electron affin- The degradation of pyrimidine groups at the temperature ity of pyridinic nitrogen dopants at the edge of GNRs are expected range expected for the dehydrogenative interibbon coupling in- to inductively downshift the frontier electronic bands,[2,46,47] thus hibits the formation of N-doped NPG by using monomer 3. In spectroscopic measurements represent a further evidence of order to facilitate the lateral coupling at lower temperatures and their presence. Indeed, the comparative dI/dV spectra acquired preserve the integrity of pyrimidine rings, we have rationally de- at the undoped and N-doped GNRs displayed in Figure 4a exhibit signed monomer 4, already introduced in Figure 1. The new a downshift of the valence band maximum (VBM) and conduc- molecular precursor contains chlorine atoms at position 2 of the tion band minimun (CBM) of 0.3 and 0.2 eV respectively, with pyrimidine rings, which is otherwise hydrogenated in monomer a minor increase in the gap from 1.25 to 1.35 eV. The identifi- 3. The presence of chlorine is expected to decrease the tem- cation of the VBM and CBM is supported by the comparison of perature of interibbon coupling by replacing the dehydrogena- constant height dI/dV maps with the corresponding wave func- tive path with a second Ullmann coupling reaction. However, tions, as obtained from our ab initio calculations, both showing we note that, when reducing the cross-coupling temperature, the same nodal symmetry (Figure 4b, see Section S1.4, Support- care has to be take on avoiding thermal overlap with other pro- ing Information for more details in the band identification). Ac- cesses such as the intraribbon cyclodehydrogenation, in order to cordingly, the band structures obtained in the same calculations keep a selective, sequential reaction path and avoid ill-defined reproduce the rigid downshift induced by the N heteroatoms (see structures.[49,50] Figure S17, Supporting Information for equivalent calculations When deposited at room temperature, monomer 4 behaves on the corresponding NPG structures). The slightly larger value similar to 3, forming large, ordered islands driven by intermolec- of 0.54 eV, we obtain for the shift can be explained by the ab- ular halogen⋅⋅⋅H and N⋅⋅⋅H interactions (Figure 6e). It also main- sence of the substrate in the calculations, which can partially tains the same staggered conformation, with the lateral distance screen the inductive shift observed experimentally. The equiv- between brightest lobes of 7.4 Å. The only distinct characteristic alent DFT calculations carried out for the corresponding NPG at this stage of the reaction is the increase of the apparent height structures to 2.9 Å (as compared to 2.3 Å found for 3), which we attribute to The homochiral interibbon N⋅⋅⋅H interaction configuration of the presence of the chlorine atoms at position 2 of the pyrimidine Figure 3b represents an ideal preconformation for the dehydro- rings (see Figure S16, Supporting Information). genative C-C cross-coupling required for the lateral bonding of At 100°C of postannealing, the homolythic cleavage of the C- the GNRs into a N-doped NPG, in analogy to that leading to the Br starts to take place, and finalizes before 140°C, as indicated by undoped NPG.[35] To test this final reaction step, we gradually in- the thermal evolution of the Br 3d peak (Figure 6b). During this crease the postannealing temperature above the 300°C required reaction step, the chlorinated pyrimidine rings do not seem to be for the formation of GNRs. For the temperature range between affected, according to the close to stoichiometric C:Cl:N core level 300–350°C, BR-STM images reveal how the pyrimidine groups relation. We note that in the C 1s core level, a new peak related start to degrade by following a ring opening reaction that, at the to the C-Cl (C[C2 Cl]) component appears at the highest binding temperature of 400°C, end with a major cleavage (see Figure S18, energy tail at 286.9 eV, in agreement with that found for other Supporting Information for a more detailed description of the chlorinated adsorbates.[51] Small Methods 2023, 2300768 2300768 (5 of 10) © 2023 The Authors. Small Methods published by Wiley-VCH GmbH 23669608, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/smtd.202300768 by Readcube (Labtiva Inc.), Wiley Online Library on [16/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.small-methods.com Figure 4. a) dI/dV spectra on undoped (red plot) and nitrogen-doped (blue plot) graphene nanoribbons with their respective bond-resolved STM images. The spectra is acquired at the edge of the bay region, where the wavefunction overlap with the tip is maximized.[35,44] b) Constant height dI/dV maps acquired at -0.5 and 0.9 V at the edge of a nitrogen-doped graphene nanoribbon (top), and the respective VBM and CBM wavefunctions obtained by DFT (bottom). STM parameters of the spectroscopic images are listed in Table S8 (Supporting Information). c) DFT calculations of the band structure of undoped and nitrogen-doped graphene nanoribbons, The frontier bands are highlighted by yellow lines, and the bandgap and the rigid band offset induced by nitrogen heteroatoms are indicated with dashed lines. Small Methods 2023, 2300768 2300768 (6 of 10) © 2023 The Authors. Small Methods published by Wiley-VCH GmbH 23669608, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/smtd.202300768 by Readcube (Labtiva Inc.), Wiley Online Library on [16/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.small-methods.com of the two type of polymers. For poly-4, instead of linear chains coexisting in both fcc and hcp tracks of the Au herringbone re- construction, we find smaller clusters of meandering polymer chains only in fcc regions of a distorted surface reconstruction (see Figure S19, Supporting Information). The preference to fcc regions,[52] and the removal of the reconstruction induced by the strong interaction with the surface,[53] have both been reported for alkali adsorbates. A similar chlorine-induced redistribution of the herringbone reconstruction has also been observed for another chlorinated graphene nanoribbon.[54] We therefore can conclude that the inclusion of a chlorine atom at the pyrimidine rings set interactions that have profound effects on the morphol- ogy and distribution of the corresponding GNRs. After annealing to 300°C, the absence of Br and Cl core lev- els indicates that both dechlorination and cyclodehydrogenation have occurred. The former is supported by the fact that the C 1s multiplet can be fitted by discarding the C-Cl component. The lat- ter is, on the other hand, inferred from the fact that halogen des- orption in these on-surface synthesis processes is prompted by hydrogenation with the residual hydrogen released in reactions such as the cyclodehydrogenation.[42,55] The planarization of polymer chains induced by the cyclode- hydrogenation is directly visualized by STM, as can be seen in Figure 6g. The analysis of STM images also reveals that the lack of order and formation of non-selective covalent bonds that in- hibit the formation of the long-range ordered arrays of laterally fused GNRs that are needed for the realization of NPG. Similar non-selective coupling reactions have also led to irregular struc- tures in the synthesis of armchair ribbons.[49] However, close-up BR-STM images of the series of aligned ribbon pairs that a ran- domly found on the surface reveal their lateral coupling. This observation proofs that, if the GNRs are prealigned, the hierar- chical (C-Br vs C-Cl) Ullmann strategy can be selective enough to form a nanoporous graphene structure by the formation of GNRs and their lateral coupling. It is also a direct demonstration that the reduction of 100–150°C of the cross-coupling tempera- ture achieved by replacing the dehydrogenative path by a second Ullmann coupling reaction can effectively prevent the degrada- tion of the pyrimidine side groups throughout all the reaction sequence needed for the formation of a N-doped NPG. 3. Conclusion In summary, we have systematically studied thermally induced on-surface synthetic routes to grow N-doped graphene nanorib- Figure 5. Tracking N 1s XPS core level signal during the OSS reactions of monomer 3. a) N 1s core level peaks at RT (black), 300°C (yellow) and bons and nanoporous graphene, the latter by the lateral coupling 400°C (purple). Fitting characteristics are listed in Supporting Informa- of ribbons. Particular focus has been dedicated to two aspects: i) tion. b) Relative area between the corrected signal of the N 1s peak versus the effect of the position and number of nitrogen heteroatoms the C 1s peak (blue plot) and between the corrected signal of the C 1s peak in the functional groups on the thermal stability of the precursor and the Au 4f doublet peaks. The nominal stoichiometric value has been in- and intermediates; ii) the search for reaction paths that preserve dicated with a grey dashed line. c) BR-STM images of GNRs postannealed the functional groups throughout the whole process by lowering at 300°C–350°C (left image) and at temperatures higher than 350°C (right the onset temperature for the cyclodehydrogenation and interib- image). STM parameters are listed in Table S8 (Supporting Information). bon coupling steps that lead correspondingly to the formation of GNRs and their lateral coupling into NPG. The debromination step tracked by XPS is consistent with the Regarding the first point, we have found that pyrimidine Ullmann polymerization observed by STM at 150°C. Similar to groups are more stable than pyridine substituents in this OSS that found for monomer 3, the large monomer islands rearrange transformation. The reactivity of pyridine functionalized pre- at this temperature into small polymer bunches. There are, how- cursors leads to premature unwanted polymerization prior to ever, substantial differences in the morphology and distribution sublimation, inhibiting their deposition by thermal evaporation. Small Methods 2023, 2300768 2300768 (7 of 10) © 2023 The Authors. Small Methods published by Wiley-VCH GmbH 23669608, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/smtd.202300768 by Readcube (Labtiva Inc.), Wiley Online Library on [16/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.small-methods.com Figure 6. On-surface synthesis of laterally connected edge nitrogen-doped graphene nanoribbons obtained from monomer 4. a–d) XPS spectra of C 1s, Br 3d, Cl 2p, and N 1s at different temperatures of postannealing. e–h) Large scale (top) and closeview (bottom) STM images of the on-surface synthesis of the laterally connected nitrogen doped graphene nanoribbons. e) Self-assembled islands of as-deposited molecules. f) Representative bunches of 1D polymers after postannealing to 150°C. g) Planarized and laterally connected ribbons after postannealing at 300°C. STM parameters are listed in Table S8 (Supporting Information). In contrast, the pyrimidine functionalized precursors are stable However, the generation of long-range ordered NPG structures is enough for sublimation and withstand all reaction steps up to limited by the prior irregular interconnection of GNRs via a non- the formation of GNRs, but degrade before their lateral fusion selective interibbon coupling. We note that, as recently demon- by dehydrogenative coupling. The temperature of this last step is strated, these undesired reactions can be avoided by guiding the reduced by 100–150°C as compared to the dehydrogenative cross- reaction with an interdigitated GNR template, thereby enabling coupling of the undoped GNR counterparts by the introduction the formation of an NPG with alternating undoped and N-doped of chlorine atoms at position 2 of the pyrimidine moieties, thus GNRs.[47] replacing the dehydrogenative path by a second Ullmann cou- Our study reflects the relevance of molecular precursor design pling reaction. We find that, when GNRs are properly aligned, the in the preservation of functional groups and the reduction of re- interibbon Ullmann cross-coupling can be induced selectively action steps, both key for the synthesis of heteroatom doped and at temperatures where the pyrimidine group is preserved. We functionalized graphene nanostructures, and demonstrates that thereby demonstrate that pyrimidine groups withstand all the re- a rational design can enable the generation of covalent suprarib- action sequence up to the formation of laterally coupled GNRs. bon structures such as nanoporous graphene. Small Methods 2023, 2300768 2300768 (8 of 10) © 2023 The Authors. Small Methods published by Wiley-VCH GmbH 23669608, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/smtd.202300768 by Readcube (Labtiva Inc.), Wiley Online Library on [16/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.small-methods.com 4. Experimental Section (MINECO), Grant Nos. CEX2021-001214-S, PID2019-107338RB-C65, PID2019-107338RB-C62, and PGC2018-096955-B-C43 funded by Synthesis of the Molecular Precursors: A detailed description of the syn- MCIN/AEI /10.13039/501100011033, FLAG-ERA grant LEGOCHIP thetic route for the preparation of each molecular precursor by means of Projects PCI2019-111890-2 and PCI2019-111933-2 funded by MCIN/AEI solution chemistry could be found in the Supporting Information. /10.13039/501100011033 and cofounded by the European Union, Grant Sample Preparation: The Au(111) single crystal was prepared by re- Nos. TED2021-132388B-C41 and TED2021-132388B-C42 funded by peated sputter-anneal cycles using Ar+ ions at an energy of 1 keV and an- MCIN/AEI /10.13039/501100011033 and the European Union NextGen- nealing to 470 °C. Precursors 3 and 4 were deposited at a rate of 0.013 erationEU/ PRTR, the GenCat (Grant No. 2017SGR1506), the European monolayer/min and 0.005 monolayer/min respectively with the Au(111) Union MaX Center of Excellence (EU-H2020 Grant No. 924143), and Xunta held at room temperature, where the monolayer was referred to the satu- de Galicia (Centro de Investigación de Galicia accreditation 2019–2022, rated self-assembled layer of precursors. Both monomers were sublimated ED431G 2019/03, and Oportunius Program). C.M. was supported by from the same commercial Dodecon OMBE fourfold Knudsen cell. The Grant RYC2019-028110-I funded by MICIN/AEI/10.13039/501100011033 sample temperature was measured by using a thermocouple directly in and by the European Social Fund “ESF Investing in your future”. M.T. was contact with the sample. The base pressure during evaporation was below supported by Spanish State Research Agency/FSE (ref. BES-2017-08078, 1 × 10−9 mbar. project ref. SEV-2013-0295-17-2). The authors thankfully acknowledge the X-ray Photoelectron Spectroscopy: XPS spectra were acquired with a computer resources at MareNostrum and the technical support provided SPECS Phoibos 150 hemispherical energy analyzer using a monochro- by Barcelona Supercomputing Center (RES-QCM-2019-1-0051). matic Al K𝛼 X-ray source at 1486.6 eV. Measurements after each anneal- ing step were carried out with the sample back at room temperature. XPS spectra were referenced to the Fermi level, fitted with a FermiDirac distri- bution at the sample temperature. The fitting results are summarized in Conflict of Interest Section S1.3 (Supporting Information). The authors declare no conflict of interest. Scanning Tunneling Microscopy and Spectroscopy: The topographic STM images were usually acquired in constant current (CC) mode and without any intentional functionalization of the tip apex. For intraribbon resolution, images were acquired in constant height (CH) mode, and with Data Availability Statement a CO-functionalized tip to perform bond-resolved STM (BR-STM). All con- The data that support the findings of this study are available from the cor- ditions for image acquisition were summarized in Table S8 (Supporting responding author upon reasonable request. Information). For the tip functionalization, CO molecules were first intro- duced on the Au surface by dosing the cryostat with CO gas at a pressure of 5 × 10−7 mbar for 30 s with the cryoshield doors of the STM open (tem- perature range from 5 to 15 K). In the STM images after CO dosing, CO Keywords molecules appear assemble in 2D clusters that nucleate at the elbows of graphene nanoribbons, heteroatom doping, nanoporous graphene, on- the fcc sites of the Au(111) herringbone reconstruction. The tip was func- surface syntheses, scanning tunneling microscopy, X-ray photoelectron tionalized by a gentle indentation onto one of these CO clusters, or spon- spectroscopy taneously by using a relatively small tip-sample distances (low voltages). Tip functionalization was recognized by a significant enhancement of the image resolution. The higher resolution of these tips were related to dy- Received: August 1, 2023 namic effects in the CO-metallic tip junction as it interacts with the atoms Published online: beneath.[56] dI/dV spectra of Figure 4 were acquired in CH mode, whereas the conductance maps were acquired in CC mode with a standard metal- lic tip using the lock-in technique. All the parameters are listed in Table S8 (Supporting Information). [1] X. Wang, X. Li, L. Zhang, Y. Yoon, P. K. Weber, H. Wang, J. Guo, H. DFT Calculations: All the electronic structure analysis in this work had Dai, Science 2009, 324, 768. been performed with DFT calculations using the SIESTA code.[57–59] . For [2] C. Bronner, S. Stremlau, M. Gille, F. Brauße, A. Haase, S. Hecht, P. the simulations the authors used the PBE form of the GGA exchange- cor- Tegeder, Angew. Chem., Int. Ed. 2013, 52, 4422. relation functional, with an optimized DZP basis set with diffuse orbitals [3] H. Kim, K. Lee, S. I. Woo, Y. Jung, Phys. Chem. Chem. Phys. 2011, 13, for C and H atoms and the default DZP basis set for N with an energy 17505. shift of 0.01 Ry. The structures were relaxed up to a force tolerance of 0.01 [4] M. Liu, Y. Song, S. He, W. W. Tjiu, J. Pan, Y. Y. Xia, T. Liu, ACS Appl. eV Å−1 . To avoid interactions between periodic replicas of the 2D layer, Mater. Interfaces 2014, 6, 4214. a large vacuum region ≈35 nm was included in the simulation box. The [5] X. Hou, Q. Hu, P. Zhang, J. Mi, Chem. Phys. Lett. 2016, 663, 123. mesh cut-off was set to 400 Ry. To sample the reciprocal space, they used [6] A. Yadegari, L. Samiee, S. Tasharrofi, S. Tajik, A. Rashidi, F. Shoghi, M. a Monkhorst–Pack grid with 30 k points in the longitudinal (ribbon) direc- Rasoulianboroujeni, M. Tahriri, S. J. Rowley-Neale, C. E. Banks, RSC tion and at least 3 k points in the transversal direction. Post processing of the results was done using the sisl python package.[60] Adv. 2017, 7, 55555. [7] Y. Bian, H. Wang, J. Hu, B. Liu, D. Liu, L. Dai, Carbon 2020, 162, 66. [8] C. Bie, H. Yu, B. Cheng, W. Ho, J. Fan, J. Yu, Adv. Mater. 2021, 33, 1. Supporting Information [9] E. F. Holby, C. D. Taylor, Appl. Phys. Lett. 2012, 101, 1. [10] Z. Liang, D. Yang, P. Tang, C. Zhang, J. Jacas Biendicho, Y. Zhang, J. Supporting Information is available from the Wiley Online Library or from Llorca, X. Wang, J. Li, M. Heggen, J. David, R. E. Dunin-Borkowski, Y. the author. Zhou, J. R. Morante, A. Cabot, J. Arbiol, Adv. Energy Mater. 2021, 11, 5. [11] J. Amaro-Gahete, A. M. Kaczmarek, D. Esquivel, C. Jiménez- Acknowledgements Sanchidrián, P. Van Der Voort, F. J. Romero-Salguero, Chem. - Eur. The authors acknowledge Guillaume Sauthier for assisting us with J. 2019, 25, 6823. the XPS experiments. This research was funded by the CERCA Pro- [12] D.-e. Jiang, V. R. Cooper, S. Dai, Nano Lett. 2009, 9, 12. gramme/Generalitat de Catalunya and supported by Grant No. SEV-2017- [13] O. S. Kwon, S. J. Park, J. Y. Hong, A. R. Han, J. S. Lee, J. S. Lee, J. H. 0706 funded by the Spanish Ministry of Economy and Competitiveness Oh, J. Jang, ACS Nano 2012, 6, 1486. Small Methods 2023, 2300768 2300768 (9 of 10) © 2023 The Authors. Small Methods published by Wiley-VCH GmbH 23669608, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/smtd.202300768 by Readcube (Labtiva Inc.), Wiley Online Library on [16/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.small-methods.com [14] Q. Chen, X. Yang, J. Membr. Sci. 2015, 496, 108. [40] R. Hatada, S. Flege, W. Ensinger, S. Hesse, S. Tanabe, Y. Nishimura, [15] H. Tan, X. Zhang, Z. Li, Q. Liang, J. Wu, Y. Yuan, S. Cao, J. Chen, J. K. Baba, Coatings 2020, 10, 1. Liu, H. Qiu, iScience 2021, 24, 1. [41] A. Batra, D. Cvetko, G. Kladnik, O. Adak, C. Cardoso, A. Ferretti, D. [16] H. W. Lee, H. S. Moon, J. Hur, I. T. Kim, M. S. Park, J. M. Yun, K. H. Prezzi, E. Molinari, A. Morgante, L. Venkataraman, Chem. Sci. 2014, Kim, S. G. Lee, Carbon 2017, 119, 492. 5, 4419. [17] P. Lazar, R. Mach, M. Otyepka, J. Phys. Chem. C 2019, 123, 10695. [42] C. Bronner, J. Björk, P. Tegeder, J. Phys. Chem. C 2015, 119, 486. [18] Y. Yamada, H. Tanaka, S. Kubo, S. Sato, Carbon 2021, 185, 342. [43] M. Abyazisani, J. M. MacLeod, J. Lipton-Duffin, ACS Nano 2019, 13, [19] W. Zhao, O. Höfert, K. Gotterbarm, J. Zhu, C. Papp, H.-P. Steinrück, 9270. J. Phys. Chem. C 2012, 116, 5062. [44] H. Söde, L. Talirz, O. Gröning, C. A. Pignedoli, R. Berger, X. Feng, [20] J. Yang, X. Zhou, D. Wu, X. Zhao, Z. Zhou, Adv. Mater. 2017, 29, 1. K. Müllen, R. Fasel, P. Ruffieux, Phys. Rev. B- Condens. Matter Mater. [21] P. Błoński, J. Tuček, Z. Sofer, V. Mazánek, M. Petr, M. Pumera, M. Phys. 2015, 91, 1. Otyepka, R. Zboil, J. Am. Chem. Soc. 2017, 139, 3171. [45] R. Pawlak, X. Liu, S. Ninova, P. D’Astolfo, C. Drechsel, S. Sangtarash, [22] S. Clair, D. G. de Oteyza, Chem. Rev. 2019, 119, 4717. R. Häner, S. Decurtins, H. Sadeghi, C. J. Lambert, U. Aschauer, S. X. [23] Y. Zhang, Y. Zhang, G. Li, J. Lu, X. Lin, S. Du, R. Berger, X. Feng, K. Liu, E. Meyer, J. Am. Chem. Soc. 2020, 142, 12568. Müllen, H. J. Gao, Appl. Phys. Lett. 2014, 105, 2. [46] J. Cai, C. A. Pignedoli, L. Talirz, P. Ruffieux, H. Söde, L. Liang, V. [24] T. H. Vo, U. G. E. Perera, M. Shekhirev, M. Mehdi Pour, D. A. Kunkel, Meunier, R. Berger, R. Li, X. Feng, K. Müllen, R. Fasel, Nat. Nanotech- H. Lu, A. Gruverman, E. Sutter, M. Cotlet, D. Nykypanchuk, P. Zahl, nol. 2014, 9, 896. A. Enders, A. Sinitskii, P. Sutter, Nano Lett. 2015, 15, 5770. [47] M. Tenorio, C. Moreno, P. Febrer, J. Castro-Esteban, P. Ordejón, [25] T. Marangoni, D. Haberer, D. J. Rizzo, R. R. Cloke, F. R. Fischer, Chem. D. Peña, M. Pruneda, A. Mugarza, Adv. Mater. 2022, 34, - Eur. J. 2016, 22, 13037. 2110099. [26] R. E. Blackwell, F. Zhao, E. Brooks, J. Zhu, I. Piskun, S. Wang, A. [48] L. E. Dinca, C. Fu, J. M. MacLeod, J. Lipton-Duffin, J. L. Brusso, C. Delgado, Y.-L. Lee, S. G. Louie, F. R. Fischer, Nature 2021, 600, 647. E. Szakacs, D. Ma, D. F. Perepichka, F. Rosei, ACS Nano 2013, 7, [27] P. Spijker, S. Kawai, S. Saito, A. S. Foster, E. Meyer, S. Osumi, S. 1652. Yamaguchi, Nat. Commun. 2015, 6, 1. [49] P. H. Jacobse, A. van den Hoogenband, M. E. Moret, R. J. Klein [28] R. R. Cloke, T. Marangoni, G. D. Nguyen, T. Joshi, D. J. Rizzo, C. Gebbink, I. Swart, Angew. Chem., Int. Ed. 2016, 55, 13052. Bronner, T. Cao, S. G. Louie, M. F. Crommie, F. R. Fischer, J. Am. [50] P. H. Jacobse, K. A. Simonov, M. J. Mangnus, G. I. Svirskiy, A. Chem. Soc. 2015, 137, 8872. V. Generalov, A. S. Vinogradov, A. Sandell, N. Mårtensson, A. B. [29] G. D. Nguyen, F. M. Toma, T. Cao, Z. Pedramrazi, C. Chen, D. J. Rizzo, Preobrajenski, I. Swart, J. Phys. Chem. C 2019, 123, 8892. T. Joshi, C. Bronner, Y. C. Chen, M. Favaro, S. G. Louie, F. R. Fischer, [51] M. R. Ajayakumar, C. Moreno, I. Alcón, F. Illas, C. Rovira, J. Veciana, M. F. Crommie, J. Phys. Chem. C 2016, 120, 2684. S. T. Bromley, A. Mugarza, M. Mas-Torrent, J. Phys. Chem. Lett. 2020, [30] J. Li, P. Brandimarte, M. Vilas-Varela, N. Merino-Díez, C. Moreno, A. 11, 3897. Mugarza, J. S. Mollejo, D. Sánchez-Portal, D. Garcia De Oteyza, M. [52] F. Hanke, J. Björk, Phys. Rev. B- Condens. Matter Mater. Phys. 2013, 87, Corso, A. Garcia-Lekue, D. Peña, J. I. Pascual, ACS Nano 2020, 14, 1. 1895. [53] W. Gao, T. A. Baker, L. Zhou, D. S. Pinnaduwage, E. Kaxiras, C. M. [31] M. Panighel, S. Quiroga, P. Brandimarte, C. Moreno, A. Garcia-Lekue, Friend, J. Am. Chem. Soc. 2008, 130, 3560. D. Rey, G. Sauthier, G. Ceballos, D. Peña, A. Mugarza, ACS Nano [54] R. Zuzak, P. Brandimarte, P. Olszowski, I. Izydorczyk, M. 2020, 14, 11120. Markoulides, B. Such, M. Kolmer, M. Szymonski, A. Garcia- [32] L. Talirz, P. Ruffieux, R. Fasel, Adv. Mater. 2016, 28, 6222. Lekue, D. Sánchez-Portal, A. Gourdon, S. Godlewski, J. Phys. Chem. [33] Y. F. Zhang, Y. Zhang, G. Li, J. Lu, Y. Que, H. Chen, R. Berger, X. Feng, Lett. 2020, 11, 10290. K. Müllen, X. Lin, Y. Y. Zhang, S. Du, S. T. Pantelides, H. J. Gao, Nano [55] A. Mairena, M. Baljozovic, M. Kawecki, K. Grenader, M. Wienke, K. Res. 2017, 10, 3377. Martin, L. Bernard, N. Avarvari, A. Terfort, K.-H. Ernst, C. Wäckerlin, [34] M. Ohtomo, H. Hayashi, K. Hayashi, H. Jippo, J. Zhu, R. Hayashi, J. Chem. Sci. 2019, 10, 2998. Yamaguchi, M. Ohfuchi, H. Yamada, S. Sato, ChemPhysChem 2019, [56] P. Jelinek, J. Phys.: Condens. Matter 2017, 29, 34. 20, 3366. [57] E. Artacho, E. Anglada, O. Diéguez, J. D. Gale, A. García, J. Junquera, [35] C. Moreno, M. Vilas-Varela, B. Kretz, A. Garcia-Lekue, M. V. Costache, R. M. Martin, P. Ordejón, J. M. Pruneda, D. Sánchez-Portal, J. M. M. Paradinas, M. Panighel, G. Ceballos, S. O. Valenzuela, D. Peña, A. Soler, J. Phys.: Condens. Matter 2008, 20, 6. Mugarza, Science 2018, 360, 199. [58] A. García, N. Papior, A. Akhtar, E. Artacho, V. Blum, E. Bosoni, P. [36] C. Moreno, M. Panighel, M. Vilas-Varela, G. Sauthier, M. Tenorio, G. Brandimarte, M. Brandbyge, J. I. Cerdá, F. Corsetti, R. Cuadrado, V. Ceballos, D. Peña, A. Mugarza, Chem. Mater. 2019, 31, 331. Dikan, J. Ferrer, J. Gale, P. García-Fernández, V. M. García-Suárez, [37] C. Moreno, M. Paradinas, M. Vilas-Varela, M. Panighel, G. Ceballos, S. García, G. Huhs, S. Illera, R. Korytár, P. Koval, I. Lebedeva, D. Peña, A. Mugarza, Chem. Commun. 2018, 54, 9402. L. Lin, P. López-Tarifa, S. G. Mayo, S. Mohr, P. Ordejón, A. [38] J. Cai, P. Ruffieux, R. Jaafar, M. Bieri, T. Braun, S. Blankenburg, M. Postnikov, Y. Pouillon, M. Pruneda, et al., J. Chem. Phys. 2020, 152, Muoth, A. P. Seitsonen, M. Saleh, X. Feng, K. Müllen, R. Fasel, Nature 204108. 2010, 466, 470. [59] M. Soler, E. Artacho, J. D. Gale, A. Garc, J. Junquera, P. Ordej, S. [39] G. Gabka, P. Bujak, K. Giedyk, K. Kotwica, A. Ostrowski, K. Daniel, J. Phys.: Condens. Matter 2002, 14, 2745. Malinowska, W. Lisowski, J. W. Sobczak, A. Pron, Phys. Chem. Chem. [60] N. Papior, sisl: v0.11.0, 2021, https://doi.org/10.5281/zenodo. Phys. 2014, 16, 23082. 597181. Small Methods 2023, 2300768 2300768 (10 of 10) © 2023 The Authors. Small Methods published by Wiley-VCH GmbH