Bicyclic Phenyl–Ethynyl Architectures: Synthesis of a 1,4‐Bis(phenylbuta‐1,3‐diyn‐1‐yl) Benzene Banister

Abstract The novel diacetylene bridged terphenylic macrocycle 1 is presented and discussed in the context of rotationally restricted “Geländer” oligomers. The 1,4‐bis(phenylbuta‐1,3‐diyn‐1‐yl) benzene bridge of diacetylene 1 is significantly longer than its terphenyl backbone, forcing the bridge to bend around the central pylon. The synthesis of molecule 1 is based to a large extent on acetylene scaffolding strategies, profiting from orthogonal alkyne protection groups to close both macrocyclic subunits by oxidative acetylene coupling sequentially. The spatial arrangement and the dynamic enantiomerization process of the bicyclic target structure 1 are analyzed. In‐depth NMR investigations not only reveal an unexpected spatial arrangement with both oligomer strands bent alongside the backbone, but also display the limited stability of the model compound in the presence of molecular oxygen.


Introduction
The fascination for conjugated macrocycles stems not only from the combination of structuralb eauty with well-defined shape and size, but also from their intrinsic physical (optical, electronic) properties. [1][2][3] The synthetic challenges pertaining to shape-persistent macrocycless uch as cyclynesa nd arenecyclynes also make them attractive from amethodsd evelopment viewpoint. [4,5] Ap articularly beautiful example is the entirely sp-hybridized macrocyclic carbona llotrope reported by Anderson and co-workersr ecently. [6] While most macrocycles aref ormally achiral (without achiralcenter), their spatial arrangement may inducet opological chirality,w ith axial chirality as ap rominent exampler elated to helical structures. [7][8][9][10][11] In 1998, Vçgtle and co-workers presentedt he concept of "Geländer"m olecules as an ew class of axial helical structures. [12] "Geländer" is the German word for banister,e xplaining the intention of the design concept. As displayed in Figure 1b, a para-terphenyl structure was complemented with two additionall inkers between neighboring phenyl subunits,l ike addingabanistert oas piral staircase( see Figure 1a). These two phenyl-interlinking chains were intendedt ow rap around the terphenyl axis in helicalm anner.W hile conceptually pioneering for atropisomers, the molecular design impedes communicating chiral information between the two biphenyl junctions.C onsequently,t he optically mute meso form is often favored over the desired helical arrangementi nt hese pairs of enantiomers. [12,13] Our contribution to improved "Geländer"s ystems was to develop al adder-type oligomer where interlinked biphenyl "rungs" are interlinked by structureso fd ifferent stepsizes. As sketched in Figure 1c,t he longero ligomerw rapped aroundt he terphenyl backbonea nd the centralb iphenyl "rung"a cts as relay communicating the chirality information betweenboth biphenyl junctionsoft he backbone. [14,15] In these new "Geländer"o ligomers (Figure 1c), surprisingly low racemization barriers were observed, although they still qualify as atropisomers. According to Oki's somewhat arbitrary definition, ahalf-life of at least 1000 seconds is requiredfor isomers to be labeled as atropos. [17] Consequently,i somersd isplaying fasterr acemization are labeled as tropos. [18,19] Another challengeo bserved in the syntheses/characterizations of these "Geländer"o ligomersw as the formation of different bicyclic systems withv arying ring-sizes. [16,20] From am aterials property perspective, "Geländer"s tructures with ac onjugated banister would be particularly interesting as model compounds with electrons delocalizedo nah elical sub-unit. We thus recently reported the synthesis of macrocycle 2 (Scheme 1), the shortestm ember of an ew "Geländer" design with ah elical oligo-para-p henylene-di-ethynylenes (OPDE) banister comprising sp-ands p 2 -hybridized carbon atoms exclusively. [21] Even thoughmacrocycle 2 already displayed an enhanced reactivity of the (bent) diacetylene,i ts room temperature stability suggested the suitability of the moleculard esign for larger model compounds.
Here the synthesis and structural investigation of the next member of the series, the bicyclic system 1 (Scheme 1) is reported. The investigations allow to draw two main conclusions concerning the molecular design: i) stability of the compounds decreases with increasing number of diacetylenes or the bridge-length in the oligomer,a nd ii)the OPDE banister in 1 behaves rather like the banister of as taircase with an inserted floor ( Figure 1f)t han as piral staircase ( Figure 1a)-instead of wrapping helically aroundt he central para-terphenyl axis (Figure 1d), it bends back on the same side of the axis (Figure 1e).

Molecular design
In analogy to our earlier strategy,t wo oligomers with different lengths werei nterlinked into al adder-types tructure, forcing the longer oligomer (OPDE as banister) to wrap around the shortero ne (oligophenylenes( OP) as axis). In this work, we intend to increase electron delocalization within the banister throught he OPDE oligomer.A sd iscussed previously for monomeric macro-cycle 2, [21] strain in the diethynyl subunit should be decreased to an acceptable level by increasing the spacing between both oligomers by an additional ethynyl linker.U sing the pictureo faladder again, tolanes were considered as "rungs"insteado fb iphenyls.
Scheme1.Chemical structures of the series of macrocyclic oligomers 1 and 2,t ogether with the retrosynthetic considerations for the assembly of 1 from the building blocks B-E. nected compound 3 with rings comprising 17 and 19 carbon atoms would be expected.
Improved control over the final macrocyclizations should be possible using precursor A 2 .I ts terminala lkynes are masked pairwise with different protection groups (PGs), enabling the consecutive and controlled closing of both macrocycles and thus the exclusive formationo ft arget compound 1.I naconvergents trategy,p recursors A can be assembled from building blocks B-E by Sonogashira-Hagihara reactions. The backbones B can be obtained by Suzuki-Miyaura cross-coupling reactions. In both of these reactions, the required regioselectivity of the coupling positionc an be directed with the large difference in reactivityofdifferent halogen substituents.
The syntheses of the two terminal side chains E and C 1 are reportede lsewhere, for the assembly of 2. [21] Variationo ft he PG in C 2 requires only minor adaptions to this protocol. In the initial strategy geared towards A 1, terphenyl backbone B 1 with three different leaving groups was considered. Consequently,a pseudos ymmetric central side chain D S exposing an alkyne group for the coupling with the backbone is required. The two alkynes of D S ,apart of the banister,a re masked with the same protection group. The assembly of buildingb lock D S requires two different alkyne protectiong roupsa tm ost and should be easily availablef rom commercialp recursors. The challenge of the strategy will be the three differently reactivel eaving groups contained in structure B 1 .T he least reactive one would be substitutedl ast, thus combining its low reactivity with the most sterically demanding reaction, caused by the bulkiness of the previouslya ttachedsidec hains.
This potentially troublesome scenario is avoidedi nt he strategy towards A 2 .H ere, the terphenyl backbone B 2 is functionalized with halogen atoms only at the terminal phenyl rings, while the central phenyla lready has the alkyne group attached.T he strategy should also ease assembling the central asymmetricmiddle side chain D AS ,where again only two different protection groups for the alkynesi nt he target structure's banister are required. In the third position of D AS (R 5 )ahalogen atom enables coupling with the alkyne of B 2 .Also, the building blocks required in the assemblyo fA 2 should be availablef rom commercial buildingb locks in afew steps.
The two terminal side chain building blocks C 1 and E were both accessible as (3-cyanopropyl)diisopropylsilyl (CPDIPS) maskeda cetylenes, as described for compound 2. [21] This was an ideal protection strategy for the assembly of A 1 ,d ue to the group's stability as well as its polarity,f acilitating chromatographic isolation of protected derivatives. [28] Thus, the "pseudo symmetric" side chain 16 was synthesized as building block D S from 1,4-dibromo-2-nitrobenzene (11)i nf ives teps (Scheme 3). Amine 12 was obtained in aB Øchamp reduction in quantitative yield and was subsequently transformed into 1,4-dibromo-2-iodobenzene( 13)i naSandmeyer-type reactioni ne xcellent yields. Profitingf rom the superior reactivity of iodine substituents in Pd-catalyzed cross-coupling reactions, 2-hydroxypropyl (HOP) acetylene was introduced in aS onogashira reaction providing 14 in good yields. Comparabler eaction conditions with elevated temperature enabled the subsequent substitution of both bromines by CPDIPS-acetylenes, affording molecule 15 in excellent 97 %y ield. The central side chain building block 16 was obtained as ay ellow oil in 79 %y ield in ar etro-Favorskii reaction,b yr efluxingc ompound 15 with sodium hydroxide in ac opper-free flask.
With all the required buildingb locks at hand, the assembly of precursor A 1 for the macrocyclization was investigated (Scheme 4). First, the iodine in terphenyl 10 was replaced with acetylene E using classical Sonogashira cross-coupling condi-tions ((Ph 3 P) 2 PdCl 2 ,C uI, THF: piperidine3:1), providing 17 after one hour at room temperature in 93 %. In the subsequent, second Sonogashira reaction, the less reactive brominei n structure 17 required increased reaction temperatures (120 8C) and pure piperidine as as olvent in order to introduce side chain 16.D espite excessive screening of reaction conditions, (Ph 3 P) 2 PdCl 2 /CuI remained the best performing catalytic system,e ven though precursor 18 was isolated in only 40 % yield by size-exclusion chromatography (SEC).
Unfortunately,a ttempts towards substituting the chlorine substituent of 18 with ap henylacetylene were not successful in our hands. In av ariety of explorative model reactions, even conditions optimizedf or arylchlorinesa te levated temperature (150 8C) were not successful and it appeared, that the chlorine of 18 is challenging to be addressed by Sonogashira cross-cou-pling conditions. [29,30] All attempts to furnish compound 19 resulted either in dehalogenation or decomposition. Considering the only moderate yield in the preceding step, the second approach via structure A 2 movedi nto the focus of interest.

Synthesis II:Assembly over precursor A 2
The design of backbone B 2 profited from the experiences we collected during the initial approach. In terphenyl 29 not only am askeda cetylene is used as at hird substituent instead of a halide with limited reactivity.A lso, the sterically most difficult ortho-position of the bottom phenyli sb earing an iodine, comparatively facilitating intended couplingr eactions.
Synthesis of backbone 29 is displayed in Scheme 5, starting with introducing aC PDIPS-acetylene.T he Sonogashira reaction between 3-iodoaniline (20)a nd CPDIPS-acetylenep rovided molecule 21 in quantitative yields. In the subsequent iodination reaction [24] not only structure 22 was obtained in 85 % yield, but side products 23, 24 and 25 werei solated (in 1.9 %, 1.6 %, and 4.9 %y ield, respectively) and identified also. After screening for amine-stable Suzuki-Miyaura reaction conditions, iodo-aryl 22 was transformed into biphenyl 26 in 84 %i solated yield using dimethoxyethane DME/EtOH/H 2 O( 4:1:1) as solvent mixture, K 2 CO 3 as base, and (Ph 3 P) 2 PdCl 2 as catalyst. In aS andmeyer-type reaction the amino group of biphenyl 26 was converted to an iodine substituent,p roviding 27 in ag ood 82 % yield. Interestingly,i nitial attempts with as imilarb uildingb lock as 26 exposing aH OP-masked alkyne were low yielding,a sa ll investigated reaction conditions gave mainly the elimination product of the HOP-protection group (2-methyl-ethenyls ubstituted alkyne) as the main compound. Using similar Suzuki-Miyaura reactionconditions as before resulted in the successful assembly of the terphenyl backbone 28 from biphenyl 27 in excellent yield (97 %). The only variationw as an increase of base equivalents (5 equiv.o fK 2 CO 3 )i no rder to compensatef or the hydrochloride salt of 2-aminophenylboronic acid. Finally,a Sandmeyer-type reaction afforded terphenyl 29 as backbone buildingb lock B 2 from precursor 28.
Startingw ithc ommercial 5-bromo-2-iodoaniline (30; Scheme 6), two consecutive Sonogashira reactions allowed to introduce both acetylenes. First, the iodine in amine 30 reacted at room temperature, giving HOP-acetylene 31 in very good 89 %i solated yield. For the reactiono fm olecule 31 with CPDIPS-acetylene, similarreactionconditions but elevated temperature( 80 8C) were applied,p roviding amine 32 in 87 %i solated yield. Again, the amine 32 was converted into analogous iodide 33 in aS andmeyer-type reaction. The latter turned out to be rather challenging, as compound 33,t he middle side chain building block D AS, was isolated in only 60 %y ield.
With all necessary buildingb locks available, the assemblyo f the bicyclic target structure 1 started by attaching all three side chains (C 2 , 33 = D AS , E)t ot he terphenylb ackbone 29 with Sonogashira reactions (Scheme 7). The free acetylene C 2 was obtainedq uantitatively by treatingt he previously published HOP and CPDIPSp rotected 1,3-diethynylbenzene [21] with tetran-butylammonium fluoride (TBAF)inT HF.The different reactivity of both halogen substituents (iodinea nd bromide) of backbone 29 allowed to substitute the iodine at room temperature selectively.U sing standard conditions ((Ph 3 P) 2 PdCl 2 ,C uI, THF, piperidine), C 2 was coupled to the backbone 29,p roviding structure 34 in good 92 %y ield. The CPDIPSp rotection group was removed selectively by treating molecule 34 with TBAF in THF,g iving intermediate 35 in nearly quantitative yield. The subsequentS onogashira couplingb etween molecule 35 and aryliodide 33 turned out to be challenging due to the pronounced tendency of molecule 35 towards dimerization by oxidative homocoupling. However,t reating the glassware with concentrated sulfuric acid to remove copperr esidues and excessive degassing of the solventm ixture (THF/ piperidine:3 /1) with argon enabled the assembly of compound 36 at room temperature in excellent 97 %i solated yield, using ((Ph 3 P) 2 PdCl 2 ,CuI) as catalytic system. The third consecutive Sonogashirar eaction required elevated reactiont emperature, like already reported for the coupling between precursors 16 and 17.T reating compounds 36 and E 2 with the mentioned, usual catalystc ombination at 120 8Ci np iperidine gave structure 37 in reasonable 65 %i solated yield.
The complex phenylene-ethynylenea rchitecture 37 comprises all carbon atoms of the targets tructure, and with its two pairs of differently maskeda cetylenes, it constitutes the desired precursor A 2 enabling the consecutive closing of both macrocycles. The orthogonal nature of the acetylene protection groups even allows choosing the order of ring-closing.
For the macrocyclization of precursor 40 into the desired target structure 1,s imilar oxidative acetylene coupling conditions werea pplied as described before for the macrocyclization of diacetylene 38.H owever,f or the addition of structure 40 by the syringe pump, 2hproved to be sufficient. After stirring the reaction mixture at room temperature for another hour,m acrocycle 1 was isolated in excellent 85 %y ield as brown solid.
Characterization,c onformation and stability of bicycle 1 The bicyclic target structure 1 was characterized by 1 Ha nd 13 CNMR spectroscopy (Supporting Information, Figure 1) as well as high-resolution mass spectrometry (HR-MS).
2D NMR spectra unambiguously corroborated the topology of the central terphenyl backbonea nd the three protruding ethynyl-phenyl side arms. The connectivity of the inner alkyne carbons( C 31 ,C 32 ,C 49 ,C 50 )t ot he outer alkyne carbon atoms (C 29 ,C 30 ,C 47 ,C 48 )w ithin the buta-1,3-diyine units could, however,n ot be monitored, ast here is no suitable long-range 3 Jo r 4 J H-C coupling constant for an HMBC-type experiment. Also, naturala bundance carbon-carbon correlation experiments were not feasible due to the low amounta nd limited stability of compound 1.A ssignment of the inner alkyne carbon atoms to the four 2D-uncorrelatedr esonances in the alkyne region of the 13 C{ 1 H} NMR spectrum was achieved to as atisfying degree by comparison with DFT calculated chemicalshifts.
The analysis of the three-dimensional arrangemento ft he dissolvedb icyclic target structure 1 by 1 H-1 HNOE spectra was surprising. The intention of the molecular designw as to force the longer1 ,4-bis(phenylbuta-1,3-diyn-1-yl) benzene oligomer to wrap helically aroundt he terphenyla xes resulting in a" Geländer"t ype arrangement as sketched in Figure 2b (compound 1a). However,t he recorded NOEs are not fully consistent with structure 1a and ab ent 1,4-bis(phenylbuta-1,3-diyn-1-yl) benzene oligomer remaining on the same side of the terphenyl subunit as displayed in Figure 2c (compound 1b)s eems to be more likely.T he relative orientation of the lower four phenyl rings (A to D, c.f. Figure 2a)i sw ell defined by strong and characteristic NOE betweenH 5 and H 28 (a), as well as H 11 and H 24 (b). The fact that the NOEs from H 5 to H 8 (c)a nd H 11 (d)a re equallys trong, indicates that phenyl rings Aa nd Ba re not oriented orthogonally,b ut have ad ihedral angle of considerably less than 908.T he distances (c)a nd (d)a re 4.5 and 4.8 in the DFT structure for 1b,c ompared to 4.3 and 4.9 respectively in 1a,t hus clearly pointing towards the structure 1b depicted in Figure 2c.Asimilar pattern is observed for the relative orientation of rings Aa nd E, where equally strong NOEs are expected from H 2 to H 35  as wella si nt oluene-D 8 solutions (data not shown), so that no unambiguous assignment to either one of the structures 1a or 1b is feasible by NOE restraints.
Altogether,t he recorded NOEs mildly favor as lightly bent terphenyl backbone with not entirelyo rthogonal phenylr ings A, B, and Ea longside the substantially bent longer1 ,4-bis(phenylbuta-1,3-diyn-1-yl) benzene oligomer.T hus the dissolved arrangements ketched as 1b (Figure2c) resembles the banister of as taircase with an inserted floor (Figure 1f)a nd not the intendedh elical staircase. 13 CNMR chemical shifts depend on the bending of the acetylenic chains and analysis according to the work of Kreuzahler et al. [34] revealt hat the lower macrocyclic ring A-C-D-B is considerably more strained than the upper one (A-C-F-E)w hich is reflected in higher shiftd ifferences for the acetylenic carbons( 12.3 and 9.0 ppm vs.7 .8 and 4.7 ppm). It also corroborates that the acetylenic moietiesa tt he ends of compound 1 are more strained (12.3 and 7.8 ppm) than the ones adjacent to the central Cp henyl ring (9.0 and 4.7 ppm).
It seems,h owever,d ifficult to quantify the differencei n strain of the upper macrocycle (A-C-F-E)f or the two proposed structures 1a and 1b.
The surprising NOE analysisn ot only disqualified the molecular design,b ut also challenged our chemical intuition. Thus, a burning issue was whether or not as imulation of the structure would have been able to predict the observeds olution arrangement. Geometry optimization was performed for both arrangements 1a and 1b using the B3LYP functional, at riplezeta basis, the RIJCOSX approximation and DFT-D3BJ disper- sion correction as implemented in the ORCA release version 4.1.2. In contrastt ot he experimental data for bicycle 1 in solution, the calculations suggested the helical arrangement 1a to be more stable than 1b by almost 28 kJ mol À1 (1a: À1917.896782035197 E h ; 1b: À1917.886403451657 E h ). We estimate the interconversion between 1a and 1b to have abarrier of about 15 kcal mol À1 (0.02436632 E h ,6 3.97 KJ mol À1 , 15.29 kcal mol À1 ,s ee Supporting Information). While we had every intention to fully assess the conformation in solutionb y extendedN MR investigations and potentially XRD, we faced a very fundamental challenge in the stability of the target compound. Its eagerness to react with molecular oxygen was already observed for the macrocycle 2 as structural synthon of the target structure, but this tendency is substantially more pronounced for bicycle 1.A sd isplayed in Figure 3, the intensity of the 1 HNMR signals of 1 decreased to 54 %w ithin 48 h, even thought he solvent ([D 6 ]benzene) was saturated with argon and the capped NMR tube was additionally sealed with aT eflon strip. While cooling slowed down the decomposition of 1,itw as not ablet oprevent it.
To identify the decomposition products,the partially degraded NMR sample was separated by SEC. Twom ain peaks were detected, from which the one with the longerr etention time was identified ast he parent bicycle 1 (2.4 mg, 3.84 mmol, 12 %). The peak with the shorter retention time (4.6mg) gave as ignal in the high-resolution mass spectrometer of m/z = 1276.33, which corresponds to the molecular formula C 100 H 44 O 2 ,e xpected for the oxygen triggered dimerization of 1. The NMR spectrum of the peak agreedw itht he hypothesized mixture of three compounds with two carbonyls each. In analogy to the macrocyclic model compound 2,f or which regioselectively one of both acetylenes of the diacetylene bridge was engaged in the oxidative dimerization( purple in Scheme 8), [21] we assumes imilar preferences for 1,c onsisting of two merged molecules 2.H owever,t he two diacetylene bridges of 1 allow three different combinationst of ormd imers, yielding in am ixture of oxidative dimers as structural isomers. The bifunctionality of 1 with respectt oo xygen triggered dimerization most likely also results in larger oligo-and polymers, which explains the loss of materiald uring the analysis of the degrading NMR sample.
This hypothesis of oxidative dimers 41 as degradation products was furthers upported by diffusiono rdered spectroscopy (DOSY). The diffusion coefficient of product 1 in C 6 D 6 was determined to be 6.34(1) 10 À10 m 2 s À1 ,w hile the diffusion coefficient of 41 was with 4.28(1) 10 À10 m 2 s À1 significantly lower. With as imple model assuming spherical moieties, the volume of the later including the first shell of solvation was calculated via the Stokes Einstein equationt ob e3 .25 times the volume of 1,whichisi nreasonable agreement with the suggested dimeric structures 41.
Of particulari nterestf or "Geländer"s tructures are their racemizationb arriers. Even though the NOE-NMR investigations challenge the helical arrangement of the banister,t he suggested strongly bent alongside arrangement of both oligomer strandsr esultsi napair of enantiomers also. Thus the question concerning the activation energy involved in the molecular enantiomerization process, macroscopically observed as racemizationo ft he sample, remains valid. Freshly purified samples of 1 were thus subjected to variable-temperature (VT) HPLC on ac hiral stationary phase. (Chiralpak IA, eluent n-hexane:iPrOH, 98:2, 1.0 mL min À1 ,c olumn oven temperature: T = 15-22 8C). Over the entire temperature range,t he elution profile displayed the separation of both enantiomers as peaks, with a substantial fraction of the sample as plateau in between both peaks, indicating structuralf lexibility of macrocycle 1 and making the isolation of pure enantiomersu nder ambient conditions impossible (a representative HPLC trace is displayed as skippingr ope in the TOC graphic). To be able to estimate the racemization barriero f1,i ts elutionprofiles in the temperature range from 288 to 298 Ki ns teps of 1Kwere recorded and analyzed (DCXplorer software packages). [35,36] Ah alf-life of the enantiomerization of only t 1/2 293 K = 159 AE 1s with an activation free energy of DG°2 93 K % 86.6 AE 1.8 kJ mol À1 was obtained. In  6 ]benzene 100 %( the sample was measured under argon, in an NMRt ube,s ealed with ap lasticlid and aT eflon strip,a tr oom temperature). a) measuredo nt he day of synthesis (underA rgon)100 %i ntensity.b)after all NMR measurements were finished ( % 48 h, at roomtemperature, in the dark,undera rgon)5 4% intensity compared to the first NMR. c) Further 5days at À26 8Cint he dark, under Argon (43 %i ntensity compared to the first NMR).
Scheme8.Oxidative dimerization of two molecules 1.a)The two triple bondsi np urple are expected to be the most reactive ones. Assuming exclusively the purple triple bonds to dimerize with O 2 ,the presence of two comparablys trainedd iacetylenes in 1 yields in 3d ifferent combinations of oxidative dimers, summarized as 41.
spite of the very limited applied temperature window guaranteeing the survival of the chiralc olumn, an Eyring plot enabled to estimate the composition of the activation free energy into its enthalpy (DH e°% 75.1 AE 0.9 kJ mol À1 )a nd entropy (DS e°% À39.3 AE 3Jmol À1 K À1 )c ontributions.
Qualitative UV/VIS spectra of the target structure 1,i ts O 2triggered degradation dimers 41,a nd its macrocyclic subunit 2 are displayed in Figure 4a). Unfortunately,macrocycle 2 and bicycle 1,a sm embers of as eries of oligomers, are not recorded in the same solvent. Due to the intrinsic labilityo ft he target structure 1,i ts electronic absorption spectrum was recorded directly after separation by SEC dissolved in CHCl 3 .D uring the same process also the dimer fraction 41 was recorded. The UV/VIS spectrum of macrocycle 2,o nt he other hand, was previously recorded in CH 2 Cl 2 . [21] Again, the poor storing stability of 2 avoided the later recording of as pectrum in CHCl 3 .
Fortunately,U V/VIS spectra recorded in CH 2 Cl 2 and CHCl 3 shoulda tl east be qualitatively comparable. [37] As expected, the extended conjugated p-systemi nt he banister of 1 is reflected in al arge bathochromic shift compared to 2.W hile compound 2 has am aximum at 289 nm and two additional maximaa t326 nm and 351 nm, the bicyclic system 1 hasa maximum at 270 nm, and two additional maximaa t3 66 nm and 396 nm. To illustrate the delocalization of the p-systemi n the banister,t he HOMOso ft he two members of the series are displayed in Figure 4b-d. The UV/VIS spectrum of the dimer fraction 41 hardly displays well-definedp eaks, but ab road absorptionist ailing out to values above 450 nm.

Conclusions
In summary,b icycle 1 as ap otentially new "Geländer"-type structure with ac onjugated helical banister (Figure1a) was successfully synthesized over 14 linear steps (5.5 %o verall yield) and fully characterized by NMR and HR-MS. In the convergent synthetic strategy,b oth macrocyclic subunits were closed consecutively,a voiding structurali somers of different ring sizes. In spiteo facalculated energy penalty DE of about 28 kJ mol À1 ,t he analysiso ft he spatial arrangemento ft he dissolved target structure by NOE NMR experiments suggests an alternative arrangement with both oligomer strands bent alongside, referredt oa s1b,a nd resembling the banister of a staircasew ith an inserted floor (Figure 1f). The target structures' racemization barrier was estimated to be DG°2 93 K % 86.6 AE 1.8 kJ mol À1 based on variable-temperature HPLC experiments on ac hiral stationaryp hase. In analogy to the macrocycle 2,t he bicyclic system 1 with its strained1 ,4-bis(phenylbuta-1,3-diyn-1-yl) benzene banister displayed ap ronounced tendency to dimerize with O 2 .
The limited stability of the target structure, along with the oxidative reactivity and the observed, unexpected spatial arrangemento ft he dissolved bicyclic system, disqualifies the investigated oligomer combinations for both, extended and/or additionally functionalized "Geländer" molecules.
We are currently working on an alternative, more modular approacht o" Geländer" systems comprising ac onjugated banister.

Generalr emarks
All chemicals were directly used for the synthesis without further purification if nothing else stated. Dry solvents were used as crown cap and purchased from Acros, Aldrich, and Fluorochem. NMR solvents were obtained from CIL Cambridge Isotope Laboratories, Inc. (Andover,M A, USA) or Aldrich. All NMR experiments were performed on Bruker AvanceI II or III HD, two or four-channel NMR spectrometer operating at 400.13, 500.13, or 600.13 MHz proton frequency.T he instruments were equipped with ad irect-observe 5mmB BFO smart probe (400, 500, and 600 MHz), an indirect-detection 5mmB BI probe (500 MHz), or af ive-channel cryogenic 5mmQ CI probe (600 MHz). All probes were equipped with actively shielded z-gradients (10 A). The experiments were performed at 298 K. All chemical shifts (d)a re reported in ppm relative to the used solvent, and coupling constants (J)a re given in Hertz (Hz). The measurements are performed at room temperature. The multiplicities are written as:s= singlet, d = doublet, t = triplet, q = quartet, quint = quintet, dd = doublet of doublet, m = multiplet. DEPT-135 experiments were performed twice for samples containing terminal alkynes using INEPT delays corresponding to 1 J CH coupling constants of 145 and 200 Hz (Reported in the supporting information). Compound 1 was fully assigned using standard COSY,T OCSY, HSQC, HMBC, and NOESY (mixing time 1s)e xperiments. The diffusion coefficients were determined in an PFGSE (pulsed field gradient spin echo) diffusion experiment using ab ipolar gradient pulse sequence. [38] The diffusion time was set to 35 ms, the Eddy current time to 5ms, and the gradient length to 1.5 ms. Gradients with a smoothed square shape (SMSQ10.100) were increased linearly in 8 steps from 5t o9 5% (2.41 to 45.74 Gcm À1 ). The sigmoidal intensity decrease was fitted with at wo-parameter fit (I0 and diffusion coefficient D) with the dosy routine implemented in topspin 3.5 [Bruker Biospin GmbH, 2017].AShimadzu GC-MS-QP2010 SE gas chromatograph system, with aZ B-5HT inferno column (30 m 0.25 mm 0.25 mm), at 1mLmin À1 He-flow rate (split = 20:1) with a Shimadzu mass detector (EI 70 eV) was used. For column chromatography,S ilicaFlashR P60 from SILICYCLE was used with ap article size of 40-63 mm( 230-400 mesh). For neutral column chromatography,S ilicaFlashR P60 from SILICYCLE was used with ap article size of 40-63 mm( 230-400 mesh) and modified by adding Buffer solution pH7 (Fluka). Therefore, am ixture of 1kgo fs ilica gel and 100 mL of diluted buffer solution (1:25, buffer:H 2 O) was subjected to rotation overnight. Recycling size-exclusion chromatography (SEC) was performed with aS himadzu Prominence System equipped with SDV preparative columns from Polymer Standards Service (two Showdex columns in series, 20 600 mm each, exclusion limit:3 0000 gmol À1 )w ith chloroform as solvent. UV/Vis absorption spectra were recorded on aJ asco V-770 Spectrophotometer.T he UV/Vis spectra were measured in a1cm quartz glass cuvettes directly after the SEC purification. For HPLC, aS himadzu LC-20ATH PLC was used equipped with ad iodearray UV/Vis detector (SPD-M20A VP from Shimadzu, l = 200-600 nm) and ac olumn oven Shimadzu CTO-20AC. The used column for separation on chiral stationary phase was aC hiralpak IA, 5 mm, 4.6 250 mm, Daicel Chemical Industries Ltd. High-resolution mass spectra (HRMS) were measured with aB ruker Maxis 4G ESI-TOF instrument, aB ruker solariX spectrometer with aM ALDI source or EI spectra were measured on aW aters Micromass AutoSpec Ultima (EI-Sector). was degassed by passing argon through for af urther 5min. The dark brown suspension was stirred at room temperature for 11 h. After the reaction was completed according to TLC, the solution was diluted with EtOAc (100 mL) and washed with H 2 O( 50 mL), brine (50 mL), and dried over Na 2 SO 4 .T he organic phase was concentrated under reduced pressure and subjected to column chromatography (340 gS iO 2 ,C y:EtOAc 94:6!60:40). Compound 34 (1.93 g, 2.77 mmol, 92 %) was obtained as ac olorless foam. R f = 0.22 (Cy/EtOAc, 3:1). 1  3 mL, 1.28 mmol, 5.0 equiv) and the reaction mixture stirred at room temperature for 1h under argon. After the reaction was completed according to TLC, the solution was diluted by the addition of water (100 mL), and the aqueous phase was extracted with EtOAc (200 mL). The organic layer was washed with water (100 mL) and brine (100 mL) and dried over Na 2 SO 4 .T he organic phase was concentrated under reduced pressure and subjected to column chromatography (100 gS iO 2 ,C y:EtOAc 94:6!65:35). Compound 35 (320 mg, 621 mmol, 97 %) was obtained as white solid. vacuum and argon cycles in the sonicator for 15 min. CuCl (49 mg, 480 mmol, 15 equiv) and Cu(OAc) 2 (122 mg, 672 mmol, 21 equiv.) were added in one portion (dark green solution), and the solution was purged again with vacuum and argon cycles in the sonicator for other 10 min. To this mixture as olution of 40 (20.0 mg, 32.0 mmol, 1.0 equiv) in pyridine (3.5 mL, 2.8 cm) was added with a syringe-pump over 2.0 h( with 0.23 mm min À1 ). One hour after full addition, the reaction was completed according to TLC (1 mL reaction mixture was evaporated and mixed with HCl 1 m (1 mL) and EtOAc (1 mL)). The pyridine was removed under reduced pressure, and the green residue was dissolved in DCM (50 mL). The yellow organic phase was washed with HCl (1 m,5 0mL, light blue) and brine (50 mL), dried over Na 2 SO 4 .T he organic phase was concentrated under reduced pressure and subjected to column chromatography (10 gS iO 2  , 121.5 (1C, C Ar , C 12 ), 98.6 (1C, C alkyne ,C 13 ), 96.2 (1C, C alkyne ,C 39 ), 95.7 (1C, C alkyne ,C 15 ), 94.2 (1C, C alkyne ,C 30 ), 92.7 (1C, C alkyne ,C 16 ), 91.9 (1C, C alkyne ,C 14 ), 90.5 (1C, C alkyne ,C 29 ), 88.9 (1C, C alkyne ,C 40 ), 88.4 (1C, C alkyne ,C 47 ), 87.5 (1C, C alkyne ,C 48 ), 82.8 (1C, C alkyne ,C 50 ), 81.9 (1C, C alkyne ,C 32 ), 81.5 (1C, C alkyne , C 31 ), 80.6 ppm (1C, C alkyne ,C 49 ). The peaks of 128.6 (1C, C alkyne ,C 44 ) and 128.5 (1C, C alkyne ,C 34 )a re only visible in the DEPT-135 experiment, as the signals are overlain by C 6 D 6 .T he inner diacetylene peaks (C 31 ,C 32 ,C 49 and C 50 )c annot be assigned from the 2D-NMRs, but were assigned according to the calculations (see Supporting