Electronic Supplementary Material (ESI) for Chemical Science. This journal is © The Royal Society of Chemistry 2016 Supporting Information: Characterizing aliphatic moieties in hydrocarbons with atomic force microscopy Bruno Schuler,1, ∗ Yunlong Zhang,2, y Sara Collazos,3 Shadi Fatayer,1 Gerhard Meyer,1 Dolores Pérez,3 Enrique Guitián,3 Michael R. Harper,2 J. Douglas Kushnerick,2 Diego Peña,3, z and Leo Gross1 1IBM Research Zurich, Säumerstrasse 4, 8803 Rüschlikon, Switzerland 2ExxonMobil Research and Engineering Company, Annandale NJ 08801, United States 3Centro de Investigación en Química Biolóxica e Materiais Moleculares (CIQUS) and Departamento de Química Orgánica, Universidade de Santiago de Compostela, Santiago de Compostela 15782 Spain (Dated: 6th December 2016) S1 Abstract In this supplemental material additional scanning tunneling microscopy (STM) and atomic force microscopy (AFM) measurements as well as NMR, optical spectroscopy and gas chromatography (GC) characterization of the ve model compounds are provided. For CPNP and CHNP the synthetic route is described. CONTENTS I. STM/AFM supporting experiments S4 A. AFM simulation for CPNP and CHNP S4 B. BPI adsorption site and apparent size S5 C. Molecular orbital imaging S6 D. Conformation change by annealing or tip-induced S8 E. Tip-induced dehydrogenation S8 F. Additional molecules on the BPD and BPI samples S10 II. BPE, BPD and BPI molecules synthesis and additional characterization S12 A. General methods S12 B. Experimental details and spectroscopic data S12 III. CPNP and CHNP molecules synthesis and additional characterization S17 A. General methods S17 B. Experimental details and spectroscopic data S17 References S29 ∗ [email protected] y [email protected] z [email protected] S2 a b c 4a 5 Å 4a 5 Å 4a 5 Å -6 Δf (Hz) 2.5 -9 Δf (Hz) 5 -15 Δf (Hz) 10 d e f 4b 5 Å 4b 5 Å 4b 5 Å -4 Δf (Hz) 2 -9 Δf (Hz) 5 -27 Δf (Hz) 23 g h i 5a 5 Å 5a 5 Å 5a 5 Å -6 Δf (Hz) 6.1 -9 Δf (Hz) 5.2 -6 Δf (Hz) 6.1 j k l 5b 5 Å 5b 5 Å 5b 5 Å -6 Δf (Hz) 6.1 -9 Δf (Hz) 5.2 -6 Δf (Hz) 6.1 Figure S1. AFM simulation. Constant-height CO tip AFM simulations using the Probe Particle Model by Hapala et al. [1,2]. S3 I. STM/AFM SUPPORTING EXPERIMENTS A. AFM simulation for CPNP and CHNP We employed the Probe Particle Model by Hapala et al. [1,2] to simulate constant- height AFM images of the molecules CPNP, 4, and CHNP, 5, in their dierent respective adsorption geometries 4a, 4b, 5a and 5b. For the calculations we used standard para- meters with a CO tip, i.e., a lateral spring constant of the tip of 0:25 N/m as supported by experiment [3]. For the molecular structures we used the relaxed geometries of the free molecules using FHI-AIMS [4]. The imaging plane was parallel to the molecular plane at distance z from the perylene core of the molecules. Geometries 4a and 4b were obtained by scanning the same molecule from either side. We chose dierences of tip heights z between the images matching the relative height osets in the experimental images shown the main text. Therefore, images Figs. S1a-c can be compared to Figs. 5b-d, respectively; Figs. S1d-f can be compared to Figs. 5f-h, respectively; and Figs. S1g-h can be compared to Figs. 6b-d, respectively. Note that in the simulations there is no substrate. But the planar substrate can be expected to yield only a constant background in constant-height AFM images. Therefore, the absolute ∆f values will dier between experiment and theory, but the relative contrast can be compared. We note that for all molecules and all geometries we observe a relatively good agreement between experiment and simulation for the large and medium tip heights (the left and the center column in Fig. S1). This agreement corroborates that our assign- ment of the dierent molecular adsorption geometries. However, for the smallest tip heights (right column in Fig. S1), the contrast above the aliphatic moieties of the molecule shows signicant deviations when comparing experiment and simulations. This indicates the dif- culties of simulating AFM images above non-planar groups at small tip heights, at which tip relaxations play an important role. It is also possible that at the small tip heights the adsorption geometry of the CPNP and CHNP molecules is altered by interaction with the tip. This eect is not captured in the simulations, which take only relaxations of the tip into account. These deviations at small tip heights illustrate the importance of establishing experimental AFM data for ngerprinting of molecular moieties for their identication in a priori unknown samples. S4 a z = 0.5Å Na+ b z = -2.7Å Cl- z = -2.6Å 21° 20 Å 10 Å -8 Δf (Hz) 1 -8 Δf (Hz) 1 Figure S2. BPI adsorption site on NaCl. a CO tip AFM image of BPI on NaCl using dierent scan heights on the substrate and close to the molecule [5]. Grid intersections indicate Cl positions. b CO tip AFM image of the same molecule at a slightly larger distance to highlight the contrast modulations among methylene groups. a 2.51 Å 2.56 Å 33.13 Å 2.56 Å 1.53 Å 2.52 Å 1.54 Å 2.52 Å 1.53 Å 10 Å b 2.3 Å 2.6 Å 34.8 Å 3.7 Å 4.1 Å 10 Å -7 Δf (Hz) 0.5 Figure S3. BPI length comparison. a Density functional theory calculated geometry of the free BPI molecule. b CO tip AFM measurmeent of BPI on NaCl. The lines in a and b indicate distances between carbon pairs (red, green) or H pairs (blue). B. BPI adsorption site and apparent size BPI's long axis adsorbs 21◦ with respect to the polar NaCl direction (see Fig. S2a). We attribute the faint AFM contrast modulations along the alkyl chain (see Fig. S2b) to adsorption height dierences in response to the substrate registry. S5 a 0.2 V b 0.2 V c 0.2 V 10 Å 10 Å 10 Å 0 z (Å) 2.7 0 z (Å) 3 0 z (Å) 1.8 d -2.7 V e -2.75 V f -2.85 V HOMO 10 Å HOMO 10 Å HOMO 10 Å 0 z (Å) 2.5 0 z (Å) 2 0 z (Å) 3.5 g 2.2 V h 2.25 V i 2.35 V LUMO 10 Å LUMO 10 Å LUMO 10 Å 0 z (Å) 4 0 z (Å) 3.7 0 z (Å) 5.8 Figure S4. STM orbital imaging of BPE, BPD and BPI. a-c In-gap STM images of BPE, BPD and BPI on NaCl(2ML)/Cu(111), respectively. d-f STM image of the HOMO resonance of the same molecules. g-i STM image of the LUMO resonance of the same molecules. In Fig. S3 the calculated (for the free molecule) and experimental distances of intra- molecular features in BPI are compared. The overall length of the molecule agrees within the measurement accuracy to the calculated model. Also the HH distances between adjacent methylene groups along the molecule's long axis (Fig. S3b dark blue) are in good agreement with the calculated values. Along the short axis (Fig. S3b light blue), however, the apparent HH distances are signicantly enlarged by about 50-60% as compared to the calculated values. One can also observe an apparent widening of the chain when going from the pyrene terminus towards the chain center. Both the enlarged appearance and widening of the chain perpendicular to its long axis are attributed to CO tip bending eects [1,6,7]. C. Molecular orbital imaging For the three dipyrene molecules (BPE, BPD and BPI) the frontier molecular orbitals could be measured by STM orbital imaging [8] as shown in Fig. S4. Both the highest occupied (HOMO) and lowest unoccupied molecular orbital (LUMO) are localized on the aromatic S6 a b c d 10 Å 0 z (Å) 2 e Cu 20 Å 20 Å 30 Å 10 Å 0 z (Å) 2.3 -20 Δf (Hz) -6 0 z (Å) 2.2 -13 Δf (Hz) -6 Figure S5. BPI post annealing to room temperature. a,b STM (a) and AFM (b) image of three BPI molecules in a conglomerate. c STM image of three BPI molecules adsorbed along a Cu(111) screw dislocation. d,e STM (d) and AFM (e) image of two BPI molecules adsorbed along a Cu(111) step edge. a b c d Cu 10 Å Cu 10 Å NaCl(2ML)/Cu 10 Å NaCl(2ML)/Cu 10 Å 0 z (Å) 2.8 0 z (Å) 2.8 0 z (Å) 9 Å2.7 0 z (Å) 2.7 e f g h Cu 10 Å Cu 10 Å NaCl 10 Å NaCl 10 Å -7 Δf (Hz) -1 -7 Δf (Hz) 3 -6.5 Δf (Hz) 2 -6.5 Δf (Hz) 2 Figure S6. BPE conformations. a,e STM (a) and AFM (e) image of BPE in the trans, at conformation on Cu. b,f STM (b) and AFM (f) image of BPE in the cis, up-down conformation on Cu. c,g STM (c) and AFM (g) image of BPE in the trans, down-up conformation on NaCl. d,h STM (d) and AFM (h) image of BPE in the trans, up-down conformation on NaCl. pyrene moieties and the aliphatic chain linker does not contribute, as expected.