Non‐Planar Structures of Sterically Overcrowded Trialkylamines

Abstract Several amines with three bulky alkyl groups at the nitrogen atom, which exceed the steric crowding of triisopropylamine significantly, were synthesized, mainly by treating N‐chlorodialkylamines with Grignard reagents. In six cases, namely tert‐butyldiisopropylamine, 1‐adamantyl‐tert‐butylisopropylamine, di‐1‐adamantylamines with an additional N‐cyclohexyl or N‐exo‐2‐norbonyl substituent, as well as 2,2,6,6‐tetramethylpiperidine derivatives with N‐cyclohexyl or N‐neopentyl groups, appropriate single crystals were generated that enabled X‐ray diffraction studies and analysis of the molecular structures. The four noncyclic amines adopt triskele‐like conformations, and the sum of the three C−N−C angles is always in the range of 351.1° to 352.4°. Consequently, these amines proved to be structurally significantly flatter than trialkylamines without steric congestion, which is also signalized by the smaller heights of the NC3 pyramids (0.241–0.259 Å). There is no clear correlation between the heights of these pyramids and the degree of the steric crowding in the new amines, presumably because steric repulsion is partly compensated by dispersion interaction. In the cases of the two heterocyclic amines, the steric stress is smaller, and the molecular structures include quite different conformations. Quantum chemical calculations led to precise gas‐phase structures of the sterically overcrowded trialkylamines exhibiting heights of the NC3 pyramids and preferred molecular conformers which are similar to those resulting from the X‐ray studies.


Introduction
The amine entityb elongs to the mosti mportant functional groups in chemistry. [1] Quite distinct classes of amines are in the center of interestd ependingo nt he different properties and the respective fields of application. Tertiary and secondary amines with sterically demanding alkyl substituents,s uch as Hünig's base 1 [2] and 2,2,6,6-tetramethylpiperidine (2), [3] play a majorrole as Brønstedbases with low nucleophilicity or as precursors of persistentn itroxylr adicals,l ike 3, [4] whicha re used for spin labeling tools ( Figure 1). [5] Many heterocyclic compounds of type 4 and free radicals derived from thesep iperidines serve as polymerization inhibitors and photostabilizers, well known as hindered amine light stabilizers (HALS). [6] Similar substances werer ecently introduced as base in frustrated Lewis pairs. [7] Furthermore, amines with bulky alkyl groups were studied in view of their pharmacological activity. [8] Finally, some sterically hindered amines have come into industrial use in gas-treatingp rocesses. [9] In academia, otherf eatures than applications, such as new records of steric congestion [10] and molecular structures of trialkylamines, are often in the focus of attention.E specially,t he question whether amines with three bulky alkyl groups will adoptaplanar insteado fthe pyramidal nitrogen atom to mini- mize the interaction of the amine substituents, has held ac ertain fascination for chemists. [11] This questionw as mainly investigated by analyzing triisopropylamine (5)s ince otherr epresentatives with significantly highers teric distressw ere not accessibleu pto quite recently ( Figure 2). Based on the results of electron diffraction studies, 5 should be very nearly planar about nitrogen with av alue for the CÀNÀCb ond angle of 119.2(3)8.T hus, as um of the angles at the nitrogen atom of 357.68 was detected, which is quite close to 3608,a ni ndication of ap lanar structure. [12] Thiso utcome was claimed to be confirmed by NMR investigations. [13] However,l ow-temperature single-crystal X-ray diffractiono f5 led to the CÀNÀCa ngle of 116.2(1)8 and as um of angles of 348.68;t husaheight of the NC 3 pyramid (nitrogen at the top) of 0.27-0.29 (depending on temperature) was determined. [14] The latter value is more than ah alf of the corresponding height in triethylamine (6), which amounts to 0.467 .I nt he solids tate, the molecule of 5 obviously adopts as omewhatf latter pyramid insteado fa planar structure, [14] and this is significantly different to the corresponding resultso ft he gas-phase electron diffraction. [12] It might be arguedt hat crystal field effects are possibly responsible for the non-planar molecular structure of 5 in ac rystallized solid.
Herein,w ed escribe the synthesis of severalt rialkylamines, which include significantly higher steric crowding than 5.I n the six cases of amines 8a and 8e-i,g eneration of singlec rystals and X-ray diffraction were successful to analyze the molecular structures. Based on theser esults, quantum chemical calculations led to precise gas-phase structures of the title compounds.

Synthesis of the amines
We mainly prepared tertiarya mines 8,i nw hich steric distress surpasses that of the standard compound 5 distinctly,b yt reating N-chlorodialkylamines 7 with Grignardr eagents in the presenceo famajor excess of tetramethylethylenediamine (TMEDA). [10a, 15] The substrates 7 were easily available by chlorination of the corresponding secondary amines witht he help of N-chlorosuccinimide. [16] Moderate yields of the alkylation products 8 were achieved as depictedi nT able 1; [17] however,i n some cases,a lternative methods led to better yields of the desired amines. For example, target compound 8a was also accessibleb yt ransforming formamide 9 into the respective chloroiminium chloride with the aid of oxalyl chloride, followed by the reaction with two equivalents of methylmagnesium bromide( Scheme 1). Furthermore, in situ generation of 1-adamantyl triflate [18] by exposure of bromide 10 to silver triflate and subsequent treatment with tert-butylisopropylamine yielded the desired product 8e. [19] Crystala nd molecular structures of the amines In the case of 8a,b,d,j,the title compoundsp roved to be colorless liquidsa tr oom temperature, whereas highly viscous liquids or waxy solids were obtained in other cases, and crystalline solids resulted by handlingo f8e-h in methanol at different temperatures. Owing to the low meltingp oints and the tendency to form plastic/disordered crystals, the crystallization and subsequent data collection as well as structure solution and refinement werev ery challenging for the compounds 8a, 8e, 8f, 8g, 8h,a nd 8i.U nfortunately,a ll attempts to obtain crystalsw hich were suitable for structure determinationw ith atomic resolution failed for compounds 8b, 8c, 8d,a nd the known [20] model compound tri-tert-butylmethanol.T he structure solution and refinement were performed using the pro-gramsS HELXS, [21] SHELXT [22] or Superflip [23] (for structure solution) and SHELXL [24] (for refinement) embedded in Olex2. [25] Thus, single-crystal X-ray studies were successful for 8a and the five amines 8e-i,s ome crystallographic details are given in Ta ble 2( comprehensive data [16,26] ).  The molecular structure of amine 8a in the single crystal shows at riskele-like conformationw ith three arms around the nitrogen atom [N1-C1-C4, N1-C5-H(C5), and N1-C8-H(C8), see Figure 3],w hich obviously enables optimal packing of the bulky alkyl groups. Similar triskele-like conformations werea lso detected in the case of the three other noncyclic amines 8e, 8f,a nd 8g (Figure 4). These preferred conformations can be characterized by selected, roughlya ntiperiplanar torsion angles as depicted in Ta ble 3. When the CÀNb ond lengths of 8a are compared with those of the sterically more stressed amines 8e, 8f,a nd 8g,s lightly greater values are found in the latter cases,e ven for CÀNb onds connectingt he same alkyl group with nitrogen;f or example, tBu-N in 8a leads to ab ond length of 1.4786(10) ,w hereas 8e revealed 1.491(3) .T he greatest CÀNd istances were always experimentally observed for the nitrogen-attached 1-adamantyl units (1.50-1.52 ). The CÀNÀCb ond angles rangedf rom 109.7 to 123.28 (Table 3). As expected, as mall angle was detected for the iPrÀNÀiPr group in 8a [113.17(6)8], while the greatest angle value resulted for amine 8f,i nw hich nitrogen is bridging two 1-adamantyl moi-eties. However,t wo quite different CÀNÀCa ngles were found in single amines even if the nitrogen is connected with ap air of the same alkyl groups.F or example, the molecular structure of 8f includes two C 1-adamantyl ÀNÀC cyclohexyl angles of 109.70(13) and 119.49(13)8;i nt he case of the smaller angle, the H(C1)ÀC1 Table 2. Some crystallographicd etails of amines 8a and 8e-i. [16,26]      The ellipsoids are shown at the 50 %p robability level. [27] bond (triskele arm) points to a1 -adamantyl unit, and obviously this is sterically more favorable than the interaction of C2 and C6 of the cyclohexyl group with the C7-C9 moiety (another triskele arm) of the other 1-adamantyl substituent (Figure4). A similar situation was detected in the molecular structure of 8g indicating two rather distinct C 1-adamantyl ÀNÀC 2-norbornyl angles of 111.1(4) and 120.4(4)8. Even thought he CÀNÀCb ond angles vary significantly,t he sum of the three angles at the nitrogen atom is alwaysi nt he range of 351.1 to 352.48 in the cases of the noncyclic amines 8a, 8e, 8f,and 8g.Hence, the four amines provedtobestructurally flatter than triisopropylamine (5), which is also demonstratedb yt he smaller heights of the NC 3 pyramids (0.241-0.259 )a ss hown in Ta ble 3. However,t here is obviously no clear correlationb etween the heights of these pyramids and the increasing degree of the steric stress in the order 8a, 8e, 8f, 8g.T his result may lead to the assumption that there is a limit in the heighto fN C 3 pyramids, which cannot be significantly smallert han 0.24 even in the case of sterically overcrowded trialkylamines.
Since product 8g was prepared from 7 and an equilibrating mixture of exo-a nd endo-2-norbornylmagnesium bromide [28] ( Table 1), an alternative stereoisomeric structure of this tertiary amine including an endo-2-norbornylg roup was also thinkable. Thus, the X-ray crystal structurea nalysis confirmed now the exo-2-norbornyl structure of 8g,w hich was previously assigned by NMRs pectroscopy. [10a] Recently,t wo non-equivalent rotamers of 8e were detected in a5 :1 ratiou sing high resolution NMR methods. Becauset hisa mine bearst hree differentb ulky alkyl groups,i to bviously is able to adoptt wo distinct triskele-like conformationsi ns olution. [10a] Them ainr otamer in solution corresponds to them olecular structureo f8ei nt he single crystal.  [27] Table 3. Some molecular details of amines 8a and 8e-i resulting from X-raystudies. [16,26] (3) 1.490 (2) 1.500 (2) 1.518 (2) 1.501 (7) 1.508 (6) 1.484 (7) 1.492 (4)   C1-N1-C6-C9 À169.93(13) C1-N1-C12-C11 169.08(13) [a] As econdm olecule in the asymmetric unit leadst oa nother seto fC ÀNb ond lengthsw ith 1.497(3),1 .497(3), and1 .501(3) and another set of bond angles at nitrogen with 122.11 (7), 110.00 (16), and 119.09(17)8 as well as another set of the corresponding torsiona ngles. [16] [b] As econd molecule in the asymmetric unit leads to another set of CÀNb ond lengths with 1.475(2), 1.490(2),a nd 1.499(2) and another set of bonda ngles at nitrogen with 114.16 (12), 113.56(13), and 116.65(12)8 as well as another set of the corresponding torsion angles. [16] In the molecular structures of the heterocyclic amines 8h and 8i,d etermined by the X-ray crystal structure analysis, the piperidine rings and also the cyclohexaner ing of 8h adopt chair conformations (Figure5). The amino group at the cyclohexane moiety of 8h is in an equatorial position, and the same is true fort he cyclohexyl and the neopentyl groups at the piperidine units of 8h and 8i,respectively.These equatorial positions are confirmedb yr oughly antiperiplanar torsion angles as depicted in Ta ble 3. However,t he conformationso ft he exocyclic substituents at the nitrogen atoms are quite different: Whereas H(C1) of the cyclohexylg roup points to C7 and the angles C1-N1-C7 [110.6(7)8]a nd C1-N1-C11[ 120.4(4)8]a re rather distinct in 8h,t he molecular structure of 8i is more symmetric with similar angles C1-N1-C6 [113.96 (12) Although amine 8h bears as econdary and two tertiary alkyl groups at the nitrogen, the sterics tress is smaller than that of 8e, 8f,a nd 8g because of the piperidine ring structure, which connects both tertiarya lkyl moieties in 8h.C onsequently,t he sum of the three angles at the nitrogen atom of 8h is slightly smaller and the height of the NC 3 pyramid is somewhat greater than the corresponding values of amine 8a that includes two secondary and only as ingle tertiary alkyl group at nitrogen (Table 3). In the case of the compound 8i with ap rimary alkyl unit at the piperidine N-atom,t he sum of the three CÀNÀ Ca ngles proves to be significantly smaller than that of triisopropylamine (5); and the height of the NC 3 pyramid is considerably greatert han that of 5.
We do not believe that crystal field effects are responsible for the non-planar molecular structures of our sterically overcrowdedt rialkylamines. In order to confirm this assumption, structuralc haracterization of these amines in the gas phase, based on high-quality quantum chemical calculations, will be helpful.

Quantum chemical calculations
In order to elucidate the effecto fc rystallization further,w ed id detailed calculations on all the speciesint he gas as well as the solid crystalline phase. Furthermore, we computed other similar compounds with smaller alkyl groups.F inally,w ed id ac onformational search of all the systems 8a-8j,i no rder to see if the conformer in the periodic crystal indeed corresponds to the minimums tructure found in the gas phase.
The effect of different generalized gradienta pproximation (GGA) density functionalsi sa lso tested. In Ta ble 4, we compare the experimental crystal structures to the experimental one and its gas phase structures.
Since different GGA functionals, like BLYP,P BE, and even the average of our computed methods yield very similarr esults, we do not believe that these will change when using another, different method.T he computed crystal structures have a lower height of the NC 3 pyramid than the experimental ones, probablyd ue to temperature and zero-point effects on the cell volume. And whereas the height is even lower in the gas phase of the less crowded compounds 8a, 8e,a nd 8h,i tb ecomes larger for the gas phase of 8f, 8g,a nd 8i.S till, all of the reported heights are in the range of 0.225-0.350 ,i ndicating that this is the value to be expected for such compounds.
Less hindered trialkylamines, such as trimethylamine, usually exhibit larger NC 3 heights. For triethylamine (6)a nd tripropylamine, different conformers will of course give different values for the heights. This is illustrated in Ta ble 5, where the pyramidal NC 3 heightc an also vary widely between 0.3 and 0.45 .I n order to discuss not only the pyramidal heights, but also the Figure 5. Molecular structures of the amines 8h (top) and 8i.All ellipsoids are shownatt he 50 %p robability level. [27] Hydrogena toms are omittedf or clarity. Table 4. Effect of different methodsa nd the environment on the molecular structures 8a, 8e, 8f, 8g, 8h,a nd 8i on the height of the NC 3 pyramid (nitrogen at the top)in. exp.
DFT average [a] BLYP [30] + D3 [31] PBE [32]  [a] Averageo ft he BLYP [29] + D3, [30] optB88-vdW, [31] PBE [32] + D3, PBE + TS, [33] RPBE [34] + D3 and vdW-DF2 [35] GGA functionals. energies needed to make the molecule more planar,w ee nforced the NC 3 substructure to be in one plane;w es et the dihedral angles to zero regarding all CNC planesa nd reoptimized the minimumc onfigurations. Although this is not necessarily the transition state on the potential energy surfaces ince all side groups will have to invert as well, it yields an estimate of the umbrella motion for the minimum structure to become planar.F or trimethylamine, the B3LYP [36] + D3/TZVPPD [37] barrier is rather high with as much as 31.2 kJ mol À1 ,w hereas it is lowered to 15 kJ mol À1 for triethylaminea nd 15.9 kJ mol À1 for tri-npropylamine. The optimized B3LYP/TZVPPD structures without dispersion yield 29.1, 17.9, and 15.2 kJ mol À1 ,r espectively,i mplying that the barrier for trimethylamine and tripropylamine are somewhat lowered by van der Waalsi nteractions, whereas for triethylamine,i ti sl arger. Triisopropylamine (5), which has been previously mentioned and investigated more than 20 years ago, is ap articulari nteresting case:Asimilar analysis between the planar and the non-planar structure in the gas phaseo ft his compound yields an extremely low inversion barriero fo nly 2.4 kJ mol À1 and a pyramidal height of only 0.200 for B3LYP + D3/TZVPPD. Whereasb asis set limit CCSD(T) [16] increases thisb arrier to 4.8-5.3 kJ mol À1 depending on the geometry used, the zero-point energy contribution lowers the barrierb y3 .0 kJ mol À1 .T hus, including the zero-point energy contribution and using more accurate post-Hartree-Fock methods, we would end up around 1.8-2.3 kJ mol À1 energy difference between the planar and the non-planar structure. The transition state has an extremely small imaginary frequency of 88 cm À1 when using B3LYP + D3, When neglectingd ispersion, the value of 2.4 kJ mol À1 decreasest he B3LYP barrier by 1.4 kJ mol À1 to 1.0 kJ mol À1 .S ince in am olecular crystal,t he dispersion is more uniform than for as ingle molecule, this decrease may explain that we obtain an almostp lanar structure with an heighto f0 .03 ,r ather independento nt he functional used when reoptimizing the solid crystal structure of this compound which has been reported as disordered with apyramidal height of 0.291 . [14] Interestingly,t he calculations provide exactly the opposite results than experiment, in which the gas phase structure was determined to be planar, [12] whereas the crystal structure was non-planar:For the gas phase, this is likely due to the extremely flat potential energy surfacea round the minimums tructure. For the solid phase, we were not able to discernt he exact cause of this discrepancyb etween experiment and theory:I ti s not the thermal and zero-point expansion of the cell volume, as larger cell volumes yield similar planar structures. The culprit for these differences are either the underestimation of DFT for the barrier in general or thermal motions which are not easily described by and modelled by theory.
Continuing with compounds 8a-8j synthesized,asimilar analysislike in Table 5isperformed in Ta ble6,whereas an analysis of 8g and 8i only gives one conformer within ag iven energy range of 30 kJ mol À1 .I ti si mportant to notet hat in all cases, the lowest energy structure of the gas phase is the one also found in the crystalline phase, see Figures3,F igure 4, and Figure 6w ith structureso f8a, 8e, 8f,a nd 8g.P erhaps contrary to initial intuition, the barriers to planarity and the pyrami-dal NC 3 heights increase again after being rather smallf or compound 8a.T his is an effect to the van der Waalsi nteractions of the large, bulky groups, which attract each other: When optimizing all structures without extra dispersion, using just B3LYP/TZVPPD (withoutD 3c orrection), all barriers of the more bulky compounds 8a-8j investigated are lowered and all pyramidal heights are consequently lowered.
In general, the conformational structure seems to have a very large effect on the planarity.F or example, compound 8e has one almostp lanar structure when being in ac onformer which is about 13 kJ mol À1 above the gas phase minimum structure, and 8f one conformer whichi sa bout 10 kJ mol À1 above the gas phasem inimum. In case ac rystal structure could trap one of thesec onformers in ap olymorph, we would obtain an NC 3 height close to zero. Overall, this effect thus showss tructures (in the gas phase) in am uch wider range Table 6. Gas phase conformer effects on the pyramidalN C 3 heightsi n using B3LYP + D3/TZVPPDof thef ive lowest conformers of the compounds 8a--8j.T he second row E is the same as the column" Energy"i n Ta ble 5, displaying the energy difference to the lowest conformer in kJ mol À1 .  than the different crystal structures of the synthesized compounds, ranging between 0.08 and 0.33 for the NC 3 heights.

Conclusions
In summary,t he molecular structures of our sterically overcrowded trialkylamines, which do not include any p system or hetero atom in proximity to the nitrogen atom, provedt ob e pyramidal with NC 3 heights that are significantly smaller than those of simple species such as trimethylamine or triethylamine. Te rtiary amines with heteroatom functionalities in the a or b positions were previously investigated, also by using X-ray studies, and led to nearly planar molecular structures, which were explainedb yo rbital interaction effects. [38] In our cases of trialkylamines, steric effects alone obviously cannot enforce complete planarization of the amine nitrogen.C rystalf ielde ffects are not responsible for the non-planar structureso fs uch trialkylamines because characterization in the gas phase, based on high-quality quantum chemical calculations,l ed to structural results which are similart ot hose of single-crystal X-ray diffraction analysis. On the one hand, van der Waals interactions of the bulky alkyl groups are ap lausible explanation that even record-breaking sterics tress cannote nforce complete planarization of the nitrogen atom. On the other hand, dispersion [39] plays also ar ole in the molecular structures of the title compounds.
As shownb yo ur X-ray studies as well as quantum chemical calculations for the molecules in the gas phase, the noncyclic amines 8a and 8e-g adopt triskele-like conformations, which can be utilized to interpret the corresponding temperature-dependenth igh-resolution NMR spectra. The same is true fort he quite differentc onformations of the heterocyclic amines 8h and 8i.C urrently,w ea re investigatingr otation processes within these amineswith the helpofdynamic NMR spectroscopy.F urthermore, we are trying to prepare tertiarya mines with even more steric crowding, for example, open-chain tri-tert-alkylamines, by oxidativer ing opening of unsaturated 2,2,6,6-tetramethylpiperidines and 2,2,5,5-tetramethylpyrrolidines.