An acyclic trialkylamine virtually planar at nitrogen. Some chemical consequences of nitrogen planarity

Article abstract of DOI:10.1021/jo100628v

(Figure presented) The synthesis of the exceedingly congested amine tris(1,3-dihydroxy-2-propyl)amine, 9, was achieved in 47- 51% overall yield. The nitrogen atom of 9 is virtually planar; it is 0.082 A out of the plane defined by the three attached carbons. The corresponding out-of-plane measurement is 0.282 A for triisopropylamine and ca. 0.4 A for uncongested trialkylamines. The N-C bonds of 9 are quite short, despite the steric congestion. The conjugate acid of 9 (viz., 9H⁺) is very strong: pK_a = 3.08 (cf. Et₃NH⁺ pK_a = 10.7). Comparison with suitable model compounds suggests 9 is less basic than predicted by ca. 1.5 pK_a units. The structure of 9H ⁺Cl^- was determined by X-ray crystallography. Here too, the nitrogen is severely flattened relative to ordinary ammonium cations. In 9H⁺Cl^-, the proton on the nitrogen of 9H⁺ forms three intramolecular hydrogen bonds to hydroxyl groups, i.e., a so-called trifurcated hydrogen bond. The NH ... O lengths in 9H⁺ are slightly shorter than comparable trifurcated hydrogen bonds. Cyclic voltammetry (CV) on 9 finds E_1/2^ox is 0.88 V, which is consistent with the inductive effect of the 1,3-dihydroxy-2-propyl groups attached to nitrogen. It is also observed that the electrochemical oxidation of 9 is reversible on the CV time scale. The ¹⁵N NMR chemical shift of the essentially planar nitrogen atom of 9 is discussed.

Full text of DOI:10.1021/jo100628v

pubs.acs.org/joc
An Acyclic Trialkylamine Virtually Planar at Nitrogen. Some Chemical
Consequences of Nitrogen Planarity
Yuanping Jie, Peter Livant,* Hui Li, Minmin Yang, Wei Zhu, Vince Cammarata,
Philip Almond, Tyler Sullens, Yu Qin, and Eric Bakker
Department of Chemistry and Biochemistry, Auburn University, Auburn, Alabama 36849-5312
livanpd@auburn.edu
Received April 2, 2010
The synthesis of the exceedingly congested amine tris(1,3-dihydroxy-2-propyl)amine, 9, was achieved in
˚
47- 51% overall yield. The nitrogen atom of 9is virtually planar; it is 0.082 A out of the plane defined by the
three attached carbons. The corresponding out-of-plane measurement is 0.282 A for triisopropylamine and
˚
˚
ca. 0.4 A for uncongested trialkylamines. The N-C bonds of 9are quite short, despite the steric congestion.
The conjugate acid of 9 (viz., 9Hþ) is very strong: pKa = 3.08 (cf. Et3NHþ pKa = 10.7). Comparison with
suitable model compounds suggests 9 is less basic than predicted by ca. 1.5 pKa units. The structure of
9HþCl- was determined by X-ray crystallography. Here too, the nitrogen is severely flattened relative to
ordinary ammonium cations. In 9HþCl-, the proton on the nitrogen of 9Hþ forms three intramolecular
hydrogen bonds to hydroxyl groups, i.e., a so-called trifurcated hydrogen bond. The NH O lengths in
3 3 3
9Hþ are slightly shorter than comparable trifurcated hydrogen bonds. Cyclic voltammetry (CV) on 9 finds
E1/2ox is 0.88 V, which is consistent with the inductive effect of the 1,3-dihydroxy-2-propyl groups attached
to nitrogen. It is also observed that the electrochemical oxidation of 9 is reversible on the CV time scale. The
15N NMR chemical shift of the essentially planar nitrogen atom of 9 is discussed.
˚
comparison, an ordinary trialkylamine generally has h > 0.4 A.
Introduction
5
˚
For example, for N(C2H5)3, h = 0.44 A.
Amines with planar nitrogens are, per se, neither rare nor
intriguing. However, trialkylamines, that is, those amines with
nitrogen bonded to three tetrahedral carbon atoms, are rarely
found to be planar at nitrogen; when they are, they arouse
curiosity. Triisopropylamine, 1, was thought for some time to be
a simple trialkylamine that was planar at nitrogen.1-3 However,
when its X-ray crystal structure was obtained, 1 was found to be
much flatter than an ordinary trialkylamine, but not planar.4
(We will use the word “planar” henceforth as a shorthand for
the phrase “planar at nitrogen”). The sum of C-N-C bond
angles (ΣφCNC) for a perfectly planar amine would be 360°. For
1, ΣφCNC was 348.6° (T = 84 K).4 Illustrated below is another
measure of nitrogen planarity, which we shall call “h”: it is the
distance from nitrogen to the plane defined by the three carbon
4
˚
atoms to which it is bonded. For 1, h is 0.292 A (T = 84 K). By
It is not unreasonable to posit that 1 tends toward
planarity in response to the steric demands of the bulky
isopropyl groups. It also stands to reason that 2 should be
(1) Bock, H.; Goebel, I.; Havlas, Z.; Liedle, S.; Oberhammer, H. Angew.
Chem., Int. Ed. Engl. 1991, 30, 187–190.
4472 J. Org. Chem. 2010, 75, 4472–4479
Published on Web 06/02/2010
DOI: 10.1021/jo100628v
r
2010 American Chemical Society

Jie et al.
JOCArticle
more planar than 1 (as it indeed is)6 because two hydrogens of
1 are replaced by bulkier OH groups to give 2. Replacing one
OH group of 2 by a much bulkier diisopropylamino group7
leads to 3, which in turn ought to be much more planar than 2.
˚
However, on going from 2 to 3, the change in h, -0.013 A, is
insignificant.6 Furthermore, there are examples of substan-
tially planarized trialkylamines that do not appear to have
enough substituents of adequate bulk to cause the observed
planarity. Someexamples of this, viz., 4,11 5,12 6,13 7,14 and8,15
are shown below. Therefore, one is led to a consideration of
effects other than purely steric ones to account for structural
trends in highly planar trialkylamines.
FIGURE 1. Orbital mixing diagram for planar R2N-CHA2. The π-
type interaction shown stabilizes the planar geometry more effec-
tively as ΔE decreases.
mixing diagram” for this molecule is shown in Figure 1.
The interaction of the occupied nitrogen p-orbital with the
unoccupied side chain 2a00 orbital is a stabilizing inter-
action that favors the planar form of the amine. As A is
made more electronegative, the side chain orbitals move to
lower energies while the nitrogen p-orbital does not,
resulting in a smaller ΔE. This makes the interaction
stronger and stabilizes the planar form still more.16 For
example, when A = Cl, the molecule is absolutely planar
17
˚
(h = 0.00 A).
We report herein the synthesis of tris(1,3-dihydroxy-2-
propyl)amine, 9. The steric demands of three secondary alkyl
groups bound to nitrogen17 and the electronic effect (viz.,
Figure 1, A = CH2OH) of three 1,3-dihydroxy-2-propyl
groups should conspire to dramatically flatten the nitrogen
atom of 9.
Results and Discussion
Our previous work with highly planar trialkylamines like
2, 3, and others led us to propose an electronic effect that
qmay operate in concert with steric effects in determining the
geometry at nitrogen in such cases.6 Briefly, consider a planar
amine, R2N-CHA2, where A stands for some single atom
(e.g., Cl) or atom in a group (e.g., C in CH3). An “orbital
Synthesis of 9. The synthesis of amine hexaalcohol 9 is
shown in Scheme 1. According to a published procedure,19
commercially available dihydroxyacetone dimer, 10, was
reductively aminated with NH4Cl to give amine tetraalcohol
11. In our hands, the yield of 11 was somewhat variable; in
particular, the product of monoamination, 2-aminopro-
pane-1,3-diol, 12, was a stubborn contaminant.
(2) Wong, T. C.; Collazo, L. R.; Guziec, F. S., Jr. Tetrahedron 1995, 51,
649–656.
The protection of 11 as the bis-acetonide 13 was reported
by the same authors.19 However, their yield of 74% of crude
13 (“sufficiently pure for further elaboration”)19 was higher
than we obtained by following their procedure. Yields on our
many attempts were scattered in the range of 30-70%.
Further, the crude 13 obtained by the literature method gave
poor results in the next reaction in our synthetic scheme, the
insertion of the carbenoid derived from diazo diester 15 into
the N-H bond of 13, as will be discussed shortly. Purifica-
tion of 13 according to the literature protocol engendered
a large loss of material. We modified the procedure for
(3) Anderson, J. E.; Casarini, D.; Lunazzi, L. J. Org. Chem. 1996, 61,
1290–1296.
€
(4) Boese, R.; Blaser, D.; Antipin, M. Y.; Chaplinski, V.; de Meijere, A.
Chem. Commun. (Cambridge) 1998, 781–782.
(5) Brady, S. F.; Singh, M. P.; Janso, J. E.; Clardy, J. Org. Lett. 2000, 2,
4047–4049.
(6) Yang, M.; Albrecht-Schmitt, T.; Cammarata, V.; Livant, P.; Makhanu,
D. S.; Sykora, R.; Zhu, W. J. Org. Chem. 2009, 24, 2671–2678.
(7) The conformational energy (“A-value”) of the OH group is 0.60 kcal/
mol (cyclohexane solvent),8 0.72 kcal/mol (acetone-d6 solvent),9 or 0.95 kcal/
mol ((CH3)2CHOH solvent),8 whereas that of the NMe2 group (a model for
the N(iPr)2 group) is 1.31 kcal/mol (toluene-d8 solvent)10 or 1.53 kcal/mol
(CFCl3-CDCl3 solvent).10
(8) Eliel, E. L.; Gilbert, E. C. J. Am. Chem. Soc. 1969, 91, 5487–5495.
(9) Aycard, J. P. Spectrosc. Lett. 1989, 22, 397–404.
(10) Booth, H.; Jozefowicz, M. L. J. Chem. Soc., Perkin Trans. 2 1976,
895–901.
(11) Zheng, X.-J.; Jin, L.-P.; Guo, J.-Q.; Zhuang, J.-Y.; Zhu, L.-G.; Mei,
Y.-H. Chin. J. Chem. 2001, 19, 907.
(12) Slugovc, C.; Mauthner, K.; Kacetl, M.; Mereiter, K.; Schmid, R.;
Kirchner, K. Chem.;Eur. J. 1998, 4, 2043.
(13) Qian, D.; Zeng, X.; Shi, X.; Cao, R.; Liu, L. Heteroat. Chem. 1997, 8,
517.
(14) Ametamey, S. M.; Prewo, R.; Bieri, J. H.; Heimgartner, H.; Obrecht,
J. P. Helv. Chim. Acta 1986, 69, 2013.
(16) (a) Alabugin, I. V.; Zeidan, T. A. J. Am. Chem. Soc. 2002, 124, 3175–
3185. (b) Alabugin, I. V.; Manoharan, M.; Zeidan, T. A. J. Am. Chem. Soc.
2003, 125, 14014–14031.
(17) Allenstein, E.; Schwarz, W.; Schrempf, E. Z. Anorg. Allg. Chem.
1987, 546, 107.
(18) Pischel, U.; Zhang, X.; Hellrung, B.; Haselbach, E.; Muller, P.-A.;
Nau, W. M. J. Am. Chem. Soc. 2000, 122, 2027–2034.
€
(19) Scott, D. A.; Krulle, T. M.; Finn, M.; Nash, R. J.; Winters, A. L.;
Asano, N.; Butters, T. D.; Fleet, G. W. J. Tetrahedron Lett. 1999, 40, 7581–
(15) Muller, C.; Bartsch, R.; Fischer, A.; Jones, P. G.; Schmutzler, R.
Chem. Ber. 1995, 128, 499.
7584.
J. Org. Chem. Vol. 75, No. 13, 2010 4473

Jie et al.
JOCArticle
SCHEME 1. Synthesis of 9
aReference 19.
preparation of 13 by first preparing and isolating the hydro-
chloride salt of 11 and then using 11 HCl in the reaction with
2,2-dimethoxypropane (DMP). We found that crystalliza-
tion of the product from hexane was a satisfactory purifica-
tion method; we obtained an 82% yield of 13, pure enough to
use in the carbenoid N-H insertion step.
In that regard, the Rh2(OAc)4 catalyst in the conversion of
13 to 16 is known to form strong complexes with Lewis bases
(including some amines).20 Such complexation would be
expected to inhibit the reaction of the catalyst with dimethyl
diazomalonate, 15 (“DDM”). In fact, Porter and co-workers
found that the catalytic activity of Rh2(OAc)4 was strongly
inhibited by primary amines.21 Since empirically the out-
come of the Rh2(OAc)4-catalyzed carbene N-H insertion
reaction does depend on the purity of 13, it seems reasonable
to aver that purification of 13 removes 5-amino-2,2-dimethyl-
1,3-dioxane, 14, the primary amine formed by reaction of
impurity 12 with DMP, and a compound capable of forming
a strong complex with Rh2(OAc)4. An earlier report from
our laboratory22 reported a 71% yield of 16, using 4 mol %
of the catalyst. We report now, with proper attention to
purification of 13, an 83% yield of 16 using 2 mol % catalyst.
Using a 20% excess of DDM brought the yield only to 84%.
Using a 50% excess did not further improve the yield.
Reduction of 16 by means of LiAlH4 (6:1 molar ratio of
LiAlH4 to 16) produced 17 in 93% yield. Unreacted starting
material could be recovered easily; the yield based on reco-
vered starting material was 98%.
3
Removal of the acetonide functionalities of 17 was accom-
plished with trifluoroacetic acid in aqueous THF, as pre-
vious work in our laboratory showed that the use of aqueous
HCl required long reaction times.22 Conversion of 9HþCF3-
-
CO2 to the free base was accomplished using an ion-
exchange resin. The overall yield of water-soluble 9 from
dihydroxyacetone dimer was 47-51%, depending on the
yield of the first step.
Compound 18, the hexatrimethylsilyl ether derivative of 9
could be prepared easily in 82% yield.
(20) (a) Doyle, M. P.; Colsman, M. R.; Chinn, M. S. Inorg. Chem. 1984,
23, 3684–3685. (b) Cotton, F. A.; Dirkarev, E. V.; Petrukhina, M. A.; Stiriba,
S.-E. Polyhedron 2000, 19, 1829–1835. (c) Prunchnik, F. P.; Starosta, R.;
Kowalska, M. W.; Galdecka, E.; Galdecki, Z.; Kowalski, A. J. Organomet.
Chem. 2000, 597, 20–28.
(21) (a) Murray-Rust, P.; McManus, J.; Lennon, S. P.; Porter, A. E. A.;
Rechka, J. A. J. Chem. Soc., Perkin Trans. 1 1984, 713–716. (b) Husinec, S.;
Juranic, I.; Liobera, A.; Porter, A. E. A. Synthesis 1988, 721–723.
(22) (a) Li, H. Synthesis of Possible Precursors to 10-P-5 and 10-N-5
Species. M. S. Thesis, Auburn University, Auburn, AL, August, 1999. (b) Yang,
M.; Wang, X.; Li, H.; Livant, P. J. Org. Chem. 2001, 66, 6729–6733.
(23) Wyatt, P.; Butts, C. P.; Patel, V.; Voysey, B. J. Chem. Soc., Perkin
Trans. 1 2000, 4222–4223.
Structure of 9. Figure 2 shows the structure of 9 derived
from X-ray crystallography. There is an intricate hydrogen-
bonding network. Every one of the six oxygen atoms of
9 is both a hydrogen-bond donor and a hydrogen-bond
acceptor. Two hydrogen bonds are internal: O1-H1 O2
3 3 3
and O3-H3 O4. In Figure 2, hydrogen-bond donor
3 3 3
4474 J. Org. Chem. Vol. 75, No. 13, 2010

Jie et al.
JOCArticle
FIGURE 3. Relation of N-C bond length to nitrogen planarity for
amines capable of the proposed electronic effect (orange triangles)
and amines incapable of it (blue circles). Compound 21 is 2-(N,N-
dicyclohexylamino)propane-1,3-diol, and 22 is 2-(2,2,6,6-tetra-
methyl-1-piperidinyl)propane-1,3-diol.6
of the 1,3-dihydroxy-2-propyl groups are large, obviously.
The electronic effect mentioned in the introduction is also
a contributing factor. Support for this is provided by the
finding that increased planarity at nitrogen is accompanied
by contraction of the N-C bond length (Figure 3). The ave-
FIGURE 2. Thermal ellipsoid plot of 9 at the 50% probability level.
Carbon atoms are shown in brown; other elements are labeled.
Oxygen atoms on neighboring molecules are lavender. Hydrogen
atoms bound to carbon are not shown. Hydrogen bonds are
indicated by dotted lines ( ).
˚
3 3 3
rage N-C bond length of 9, 1.454 A, puts it in the lowest
quartile of 1402 trialkylamine N-C bond lengths compiled
by Allen et al.25 The relation between flatter nitrogen and
shorter N-C bonds is stronger in those cases in which the
electronic effect can operate (i.e., those with heteroatoms at
R or β positions; orange triangles in Figure 3) than in cases
lacking this feature (blue circles in Figure 3). The π interac-
tion between the nitrogen 2p orbital and the adjacent σ*
orbital would be expected to result in a shorter N-C bond
length.
and acceptor atoms external to the molecule are shown in
lavender.
Figure 2 also makes it clear that the central nitrogen
(shown in blue) is virtually planar. For 9, the sum of C-N-C
˚
angles is 359.06° and h = 0.082 A. For comparison, consider
h values in a series of trialkylamines in which each nitrogen is
bonded to three secondary carbons, shown below. Here
again, it is clear that the nitrogen atom of 9 is exceedingly
flat.
Basicity of 9. The pKa of protonated 9, i.e., 9Hþ, was
found to be 3.08 ( 0.03 at 25 °C (eq 1).
Compared to an ordinary trialkylamine such as triethyl-
amine whose conjugate acid has a pKa of 10.70,26 9 would
appear to be an exceptionally weak base. However, the
presence in an amine of an electronegative OH group β to
nitrogen should weaken basicity, and 9 has six such OH
groups. Figure 4 plots pKa versus the number of -C-C-OH
groups attached to the nitrogen of various protonated
tertiary amines.
Although it is natural to ask why 9 is so flat, a reasonably
satisfying answer is difficult to provide. To any answer that
might be offered, several objections could be raised: (1) the
geometry at nitrogen might be affected by the H-bonding
network, (2) crystal packing forces might perturb the gas-
phase geometry in unpredictable ways.24 Even so, we believe
that the major forces at work can be classified under two
familiar headings: steric and electronic. The steric demands
(25) Allen, F. H.; Kennard, O.; Watson, D. G.; Brammer, L.; Orpen,
A. G.; Taylor, R. J. Chem. Soc. Perkin Trans. II 1987, S1–S19.
(26) Chawla, B.; Mehta, S. K. J. Phys. Chem. 1984, 88, 2650–2655.
(27) (a) Hamborg, E. S.; Versteeg, G. F. J. Chem. Eng. Data 2009, 54,
1318–1328. (b) Schwabe, K.; Graichen, W.; Spiethoff, D. Z. Phys. Chem.
(Munich) 1959, 20, 68–82.
(24) Alabugin, I. V.; Manoharan, M.; Buck, M.; Clark, R. J. J. Mol.
Struct. THEOCHEM 2007, 813, 21–27.
J. Org. Chem. Vol. 75, No. 13, 2010 4475

Jie et al.
JOCArticle
FIGURE 4. pKa of protonated tertiary amines as a function of the
number of OH groups β to nitrogen. Except for 9Hþ, data were
taken from ref 27. The least-squares line is a fit to the filled triangles.
FIGURE 5. Thermal ellipsoid plot of 9HþCl- at the 50% level.
Carbons, unlabeled, are shown in brown; other atoms are labeled.
Hydrogen atoms bound to carbon are not shown. Internal H-bonds
are shown by dashed lines. Cl- is not shown here but has close
contacts with OH hydrogen atoms on four neighboring cations,
One may extrapolate the data for protonated amines
having 0, 1, 2, and 3 -C-C-OH groups (filled triangles in
Figure 4) to a protonated tertiary amine with six such groups,
viz., 9Hþ, which should have a pKa of 4.62. The observed pKa
of 3.08 means compound 9 is less basic than predicted on
electronic grounds by roughly 1.5 pKa units. Should this
1.5 pKa unit decrease be called a “planarization effect”? The
0
0
0
˚
˚
˚
namely, Cl H2 , 2.341 A; Cl H3 , 2.436 A; Cl H4 , 2.359 A;
3 3 3 3 3 3 3 3 3
0
˚
Cl H5 , 2.222 A.
3 3 3
TABLE 1. Geometries of Protonated Acyclic Trialkylamines
av N-C av H-N-C av C-N-C no. of
-
-
pKa’s of Et3NHþClO4 and (iPr)3NHþClO4 are 7.9 and
7.0, respectively, measured in 2-methoxyethanol at 20 °C.28
If one allows that the electronic effects of an ethyl group and
an isopropyl group are not very different, these data suggest
the existence of a planarization effect on basicity. However,
at present it would be difficult to support the proposal
convincingly since electronic factors affect both planarity
and basicity. In any event, we believe it is reasonable to
suggest that the diminished basicity of 9 might be due to
difficulty in deforming the essentially planar nitrogen of 9 to
accommodate the additional ligand, a proton, in 9Hþ. For-
tuitously (i.e., by mistake), in the course of other work with 9,
we prepared 9HþCl- as a crystalline material suitable for
X-ray crystallography.
ammonium ion length (A) angle (deg) angle (deg) examplesa
˚
uncongested (i.e., ordinary)
107.5
107.0
(CH3)3NHþ
1.48
(HOCH2CH2)3NHþ 1.50
111.4
111.8
49
21
congested
((CH3)2CH)3NHþb
1.533
1.528
105.1
102.3
113.5
115.6
b
1
1
9Hþc
a
€
From the Cambridge Crystallographic Database. Bock, H.; Gobel,
I.; Bensch, W.; Solouki, B. Chem. Ber. 1994, 127, 347-351. cThis work.
An interesting feature of 9Hþ is the presence of a trifur-
cated (or “four-center”) hydrogen bond involving the Nþ-H
hydrogen (H1) and three neighboring oxygens (O4, O5, and
˚
O6). The relevant lengths are H1 O4, 2.255 A; H1 O5,
3 3 3
3 3 3
Structure of 9Hþ. A thermal ellipsoid plot of 9Hþ at the
50% probability level is shown in Figure 5. As shown in
Table 1, the nitrogen of 9Hþ is severely flattened relative to
two examples of “ordinary” ammonium ions. The 115.6°
average C-N-C angle is closer to the trigonal planar value
(120°) than it is to the tetrahedral value (109.5°).
˚
˚
2.218 A; and H1 O6, 2.181 A. The N-protonated con-
3 3 3
jugate acid of triethanolamine also exhibits an NH (OH)3
3 3 3
˚
trifurcated H-bond. In that case the NH O length is 2.33 A,
3 3 3
which is an average of X-ray data from 15 or so literature reports
of this cation.29 Trifurcated H-bonds of the NH (OH)3 type
3 3 3
are also reported for (HOOCCH2CH2)2NHþ(CH2CH2OH),
˚
23,and24, with average NH O lengths 2.29, 2.30, and 2.27 A,
respectively.30 It thus appears that the average NH O length
˚
in 9H , 2.22 A, is slightly shorter than these closely comparable
examples.
3 3 3
(28) Wieland, G.; Simchen, G. Liebigs Ann. Chem. 1985, 2178–2193.
(29) (a) Demir, S.; Yilmaz, V. T.; Harrison, W. T. A. Acta Crystallogr.,
Sect. E: Struct Rep. Online 2003, 59, o907. (b) Starova, G. L.; Frank-
Kamenestskaya, O. V.; Fundamenskii, V. S.; Semesnova, N. V.; Voronkov,
M. G. Proc. Nat. Acad. Sci. USSR 1981, 260, 888. (c) Harrison, W. T. A. Acta
Crystallogr., Sect. E: Struct Rep. Online 2003, 59, o1267. (d) van Mier,
G. P. M.; Kanters, J. A.; Poonia, N. S. Acta Crystallogr., Sect. C: Cryst.
Struct. Commun. 1988, 44, 334. (e) Bracuti, A. J. J. Crystallogr. Spectrosc.
Res. 1993, 23, 669. (f) Parkanyi, L.; Hencsei, P.; Nyulaszi, L. J. Mol. Struct.
1996, 377, 27. (g) Steiner, T.; Schreurs, A. M. M.; Lutz, M.; Kroon, J. New J.
Chem. 2001, 25, 174. (h) Mootz, D.; Brodalla, D.; Wiebcke, M. Acta
Crystallogr., Sect. C: Cryst. Struct. Commun. 1990, 46, 797. (i) Naiini,
A. H.; Pinkas, J.; Plass, W.; Young, V. G., Jr.; Verkade, J. G. Inorg. Chem.
1994, 33, 2137. (j) Lukevics, E.; Arsenyan, P.; Shestakova, I.; Domracheva,
I.; Kanepe, I.; Belyakov, S.; Popelis, J.; Pudova, O. Appl. Organomet. Chem.
2002, 16, 228. (k) Soumhi, E. H.; Assdoune, I.; Driss, A. Acta Crystallogr.,
Sect. E: Struct Rep. Online 2007, 63, o2827. (l) Wermester, N.; Lambert, O.;
Coquerel, G. CrystEngComm 2008, 10, 724.
3 3 3
þ
A wider search of the crystallographic database for
NH (O)3,NH (O)2(N), NH (O)(N)2,andNH (N)3
3 3 3
3 3 3
3 3 3
3 3 3
4476 J. Org. Chem. Vol. 75, No. 13, 2010

Jie et al.
JOCArticle
four-center hydrogen bonds revealed 37 more examples, includ-
ing N-protonated cryptands and several forms of nitrilotriacetic
acid.31-33 Although no examples of NH (O)(N)2 trifurcated
H-bonds were found, for the other three classes of trifurcated
3 3 3
hydrogen bonds, average NH X (X = O or N) bond
3 3 3
˚
˚
distances were 2.29 A (NH (O)3), 2.29 A (NH (O)2(N)),
3 3 3
3 3 3
and 2.31 A ((NH (N)3). Here again, 9Hþ exhibits shorter
˚
3 3 3
trifurcated H-bonds than found, on average, in the 37 related
cases.
Electrochemical Oxidation of 9. The one-electron oxida-
tion potential, E1/2ox, of 9 is 0.88 V, as measured by cyclic
voltammetry (CV). Our earlier CV investigation of severely
flattened trialkylamines6 found that values of E1/2ox in this class
of amines correlated poorly with degree of nitrogen planarity
but much more strongly with the net inductive effect of the three
alkyl groups attached to nitrogen, as suggested by Mann.34 In
ox
keeping with this idea, we note that a plot (not shown) of E1/2
for triisopropylamine, 1 (0.73 V),18 2-(dicyclohexylamino)-1,3-
propanediol, 20 (0.77 V),6 and 9 (0.88 V) versus the number
of 1,3-dihydroxy-2-propyl substituents borne by the central
(30) (a) Shkol’nikova, L. M.; Obodovskaya, A. E. Zh. Struckt. Khim.
€
FIGURE 6. Thermal ellipsoid plot of 9 at the 50% level, drawn to
emphasize the near-orthogonality of the nitrogen p-orbital and the
adjacent C-H bonds. Atoms of interest are in color.
1986, 27, 125. (b) Muller, E.; Burgi, H.-B. Acta Crystallogr., Sect. C: Cryst.
Struct. Commun. 1989, 45, 1400. (c) Timosheva, N. V.; Chandrasekaran,
A.; Day, R. O.; Holmes, R. R. J. Am. Chem. Soc. 2002, 124, 7035.
(d) Chandrasekaran, A.; Timosheva, N. V.; Day, R. O.; Holmes, R. R.
Inorg. Chem. 2003, 42, 3285. (e) Huang, W.; Chu, Z.; Xu, F. J. Mol. Struct.
2008, 885, 154.
nitrogen is linear, with r2 = 0.995. Little, if any, quantitative
significance should be attached to this three-point correlation!
Qualitatively, the trend is in the correct direction, namely,
toward more positive oxidation potentials with more electron-
withdrawing substituents.
(31) NH (O)3 (a) Nielson, A. J.; Shen, C.; Waters, J. M. Acta Crystal-
3 3 3
logr., Sect. C: Cryst. Struct. Commun. 2003, 59, m494. (b) Cody, V.; Hazel, J.;
Langs, D. Acta Crystallogr., Sect. B: Struct. Cryst. Cryst. Chem. 1977, 33,
905. (c) Whitlow, S. H. Acta Crystallogr., Sect. B: Struct. Cryst. Cryst. Chem.
1972, 28, 1914. (d) Sopo, H.; Sviili, J.; Valkonen, A.; Sillanpaa, R. Polyhedron
2006, 25, 1223. (e) Duesler, E. N.; Tapscott, R. E.; Garcia-Basallote, M.;
Gonzalez-Vilchez, F. Acta Crystallogr., Sect. C: Cryst. Struct. Commun.
1985, 41, 678. (f) Murugesu, M.; Clerac, R.; Anson, C. E.; Powell, A. K.
J. Phys. Chem. Solids 2004, 65, 667. (g) Shkol’nikova, L. M.; Suyarov, K. D.;
Poznyak, A. L.; Dyatlova, N. M.; Makarevich, S. S. Zh. Strukt. Khim.
(J. Struct. Chem.) 1988, 29, 113–119. (h) Sharma, C. V. K.; Clearfield, A.
J. Am. Chem. Soc. 2000, 122, 4394. (i) Cohen, S. M.; Meyer, M.; Raymond,
K. N. J. Am. Chem. Soc. 1998, 120, 6277. (j) Pisarevsky, A. P.; Yanovsky,
A. I.; Struchkov, Yu. T.; Kesel’man, Ya. G. Koord. Khim. (Coord. Chem.)
1994, 20, 35. (k) Junk, P. C.; Raston, C. L. Inorg. Chim. Acta 2004, 357, 595.
(l) Luger, P.; Buschmann, J.; Knochel, A.; Tiemann, D.; Patz, M. Acta
Crystallogr., Sect. C: Cryst. Struct. Commun. 1991, 47, 1860. (m) Farahbakhsh,
M.; Kogerler, P.; Schmidt, H.; Rehder, D. Inorg. Chem. Commun. 1998, 1, 111.
(n) Renaud, F.; Piguet, C.; Bernardinelli, G.; Hopfgartner, G.; Bunzli, J.-C. G.
Chem. Commun. 1999, 457. (o) Lakshminarayanan, P. S.; Suresh, E.; Ghosh, P.
Inorg. Chem. 2006, 45, 4372. (p) Chekhlov, A. N. Zh. Neorg. Khim. (Russ. J.
Inorg. Chem.) 2005, 50, 1656. (q) Davidovich, R. L.; Gerasimenko, A. V.;
Logvinova, V. B.; Hu, A.-Z. Zh. Neorg. Khim. (Russ. J. Inorg. Chem.) 2002, 47,
1610. (r) Haussuhl, E.; Haussuhl, S.; Tillmanns, E. Z. Kristallogr. 2004, 219, 644.
(s) Willey, G. R.; Meehan, P. R.; Rudd, M. D.; Drew, M. G. B. J. Chem. Soc.,
Dalton Trans. 1995, 3173. (t) MacGillivray, L. R.; Atwood, J. L. Chem.
Commun. 1996, 735. (u) Pramanik, A.; Bhuyan, M.; Choudhury, R.; Das, G.
J. Mol. Struct. 2008, 879, 88. (v) Chekhlov, A. N. Koord. Khim. (Coord. Chem.)
2007, 33, 947.
The oxidation of 9 is reversible, even at a sweep rate of
20 mV/s, scanning between 0 and 1.5 V. Most trialkylamines
examined by cyclic voltammetry exhibit an oxidation wave but
lack a reduction wave.34 The explanation commonly invoked
for this is that one-electron oxidation of nitrogen to an amine
radical cation, which is responsible for the oxidation peak, is
followed by rapid and irreversible loss of an R proton.35 That
this decomposition route, viz., the loss of an R-proton, is appa-
rently slow in the present case is a consequence of geometry: in
9, the C-H bonds R to nitrogen lie essentially in the nodal
plane of the nitrogen p-orbital, as is shown in Figure 6. The
dihedral angles C5-N1-C6-H6A, C6-N1-C7-H7A, and
C7-N1-C5-H5A are 6.6°, 16.0°, and 17.0° respectively. One
expects the geometry of the radical cation 9þ• to be virtually
unchanged. The near-orthogonality of all R C-H bonds to the
nitrogen p-orbital precludes stabilization of the incipient R
radical center via overlap with the adjacent p-orbital. Similar
CV behavior by triisopropylamine has been reported, and a
similar explanation offered.1,18
(32) NH (O)2(N) (a) Choquesillo-Lazarte, D.; Covelo, B.; Gonzalez-
3 3 3
Perez, J. M.; Castineiras, A.; Niclos-Gutierrez, J. Polyhedron 2002, 21, 1485.
(b) Tircso, G.; Benyei, A.; Kiraly, R.; Lazar, I.; Pal, R.; Brucher, E. Eur. J.
Inorg. Chem. 2007, 701. (c) Medrano, F.; Corona-Martinez, D. O.; Hopfl, H.;
Godoy-Alcantar, C. Supramol. Chem. 2007, 19, 621. (d) Finnen, D. C.;
Pinkerton, A. A. Acta Crystallogr., Sect. C: Cryst. Struct. Commun. 1997, 53,
1455. (e) Inoue, M. B.; Oram, P.; Inoue, M.; Fernando, Q. Inorg. Chim. Acta
1996, 248, 231. (f) Prince, P.; Watkins, S. F.; Fronczek, F. R.; Gandour,
R. D.; White, B. D.; Gokel, G. W. Acta Crystallogr., Sect. C: Cryst. Struct.
Commun. 1991, 47, 893.
15N NMR of 9. In D2O solvent, 9 exhibits one peak in its
15N NMR spectrum at -366.3 ppm (referenced to external
CH3NO2). Triethanolamine in the same solvent appears at
-352.3 ppm. Parameters that permit one to predict the 15N
chemical shift of a simple mono-, di-, or trialkylamine were
developed by Duthaler and Roberts.36 Using these, Wong,
Collazo, and Guziec2 compared predicted and observed
nitrogen chemical shifts of four sterically congested tertiary
(33) NH (N)3 (a) Zeng, X.; Coquiere, D.; Alenda, A.; Garrier, E.;
3 3 3
Prange, T.; Li, Y.; Reinaud, O.; Jabin, I. Chem.;Eur. J. 2006, 12, 6393.
(b) Schatz, M.; Becker, M.; Thaler, F.; Hampel, F.; Schindler, S.; Jacobson,
R. R.; Tyeklar, Z.; Murthy, N. N.; Ghosh, P.; Chen., Q.; Zubieta, J.; Karlin,
K. D. Inorg. Chem. 2001, 40, 2312. (c) Springborg, J.; Pretzmann, U.; Olsen,
C. E.; Sotofte, I. Acta Chem. Scand. 1998, 52, 289.
(35) (a) Kanoufi, F.; Zu, Y.; Bard, A. J. J. Phys Chem. B 2001, 105, 210–
216. (b) Nelsen, S. F.; Ippoliti, J. T. J. Am. Chem. Soc. 1986, 108, 4879–4881.
(c) Dinnocenzo, J. P.; Banach, T. E. J. Am. Chem. Soc. 1989, 111, 8646–8653.
(36) Duthaler, R. O.;Roberts, J. D.J. Am. Chem. Soc. 1978, 100, 3889–3895.
(34) (a) Mann, C. K. Anal. Chem. 1964, 36, 2424–2426. (b) Hull, L. A.;
Davis, G. T.; Rosenblatt, D. H.; Mann, C. K. J. Phys. Chem. 1969, 73, 2142–
2146.
J. Org. Chem. Vol. 75, No. 13, 2010 4477

Jie et al.
JOCArticle
amines, namely, diisopropylmethylamine, diisopropylethy-
lamine, di-tert-butylmethylamine, and triisopropylamine
(1). The agreement between predicted and observed shifts
(CH3OH solvent) was good for the first three amines (within
2 ppm), but 1 deviated significantly (ca. 11 ppm upfield of the
prediction). Wong et al. argued that this was consistent with
the first three amines being pyramidal and 1 being planar.
In the case of 9, it is a little more difficult to decide if a
chemical shift of -366.3 ppm is “expected” or anomalous.
15N NMR data for β-hydroxyamines are somewhat sparse,
and solvent effects on 15N chemical shifts can be substantial.
Nevertheless, using the published 15N chemical shifts of
β-hydroxyamines and of structurally similar non-β-OH con-
taining amines, one may crudely estimate the chemical shift
change resulting from a single N-C-CH3 f N-C-CH2OH
change, a parameter we will arbitrarily call “A”. The method
is illustrated in Table 2.
Experimental Section
Bis(1,3-dihydroxy-2-propyl)amine, 11. This compound was
prepared by the method of Fleet et al.19
Bis(4,4-dimethyl-3,5-dioxanyl)amine, 13. Secondary amine 11
(4.00 g, 24.2 mmol) was dissolved in methanol (60 mL), and
37.5% HCl solution (10 mL) was added and stirred for 2 h at rt.
The solvent was removed on a rotary evaporator. The residue
wasdissolvedinDMF(40mL). Undernitrogen, p-toluenesulfonic
acid (0.700 g, 3.60 mmol) and 2,2-dimethoxypropane (10.0 mL,
81.6 mmol) were added to the solution. The resulting clear
solution was allowed to stir overnight (at least 12 h), at which
time Et3N (0.600 mL, 4.00 mmol) was added and allowed to stir
for an additional 10 min. The mixture was concentrated in vacuo
and treated with Et3N (3.40 mL, 24.2 mmol) and EtOAc (150 mL).
The precipitate was removed by filtration, and the filtrate was
concentrated. The residue was crystallized from hexane to give
82% yield, mp 58.5-59.5 °C. The spectroscopic data were con-
sistent with those previously reported.19 1H NMR (400 MHz,
CDCl3): δ 1.41 (6 H, s), 1.42 (6 H, s), 2.75 (2 H, tt, J = 4.1, 6.8),
3.62 (4 H, dd, J = 6.8, 11.7), 3.93 (4 H, dd, J = 4.1, 11.7). 13
NMR (100 MHz, CDCl3): δ 23.2, 24.7, 50.0, 64.7, 98.4.
C
TABLE 2. Two Examples of the Calculation of A
Dimethyl 2-(N,N-Bis(4,4-dimethyl-3,5-dioxanyl)amino)-
malonate, 16. Under N2, to a solution of acetonide amine 13
(0.980 g, 4.00 mmol) in benzene (20 mL) was added 0.760 g of 1538
(4.80 mmol) and rhodium(II)acetate dimer (32.0 mg, 0.0700
mmol) at rt. The mixture was heated to reflux and continued
for 2.0-2.5 h, at which time TLC showed the absence of diazo
compound and acetonide amine. The solvent was removed in
vacuo. The residue was purified by silica gel chromatography
(EtOAc/hexane 1:3 v/v), which yielded white tertiary amine 16
(1.25 g, 83.3%), mp 63-64.5 °C (lit. 63-64.5 °C).22b Anal. Calcd
for C17H29NO8: C, 54.39; H, 7.79; N, 3.73. Found: C, 54.64; H,
7.78; N, 3.60. 1H NMR (400 MHz, CDCl3): δ 1.36, 1.40 (12H, 2s),
3.34 (2 H, m), 3.72 (4H, dd, J = 6.9, 12.0), 3.77 (6H, s), 3.96 (4H,
dd), 4.99 (1H, s). 13C NMR (100 MHz, CDCl3): δ 23.0, 24.3, 51.4,
52.6, 63.5, 64.3, 98.2, 170.3. EI-HRMS calcd for C17H29NO8
375.1893, found 375.1888.
These and seven more comparisons are collected in the
Supporting Information.36,37 The average valueof A is -7.4 (
1.5 ppm. As shown below, triisopropylamine can be trans-
formed to 9 by six N-C-CH3 f N-C-CH2OH changes,
leading to a predicted 15N chemical shift for 9 of -328.1 þ
6(A) = -372.5 ppm. The observed 15N chemical shift for 9,
-366.3 ppm, differs from the prediction by 6.2 ppm. We
consider this to be within the limits of uncertainty arising from
the small numberofcomparisonsused to calculatethe average
A value and the necessity of comparing chemical shifts of
compounds measured as neat samples with shifts of other
compounds measured in methanol or cyclohexane solution.
2-(N,N-Bis(4,4-dimethyl-3,5-dioxanyl)amino)-1,3-propanediol,
17. Tertiary amine 16 (0.380 g, 1.00 mmol) was dissolved in THF
(4 mL) and added dropwise to a suspension of LiAlH4 (0.230 g,
6.00 mmol) in THF (10 mL). The reaction was stirred overnight
at room temperature. Water (0.230 mL), 15% sodium hydroxide
(0.230 mL), and water (3 ꢀ 0.230 mL) were added sequentially.
The mixture was filtered, and the filtrate was evaporated. The
residue was purified by silica gel chromatography (ethyl acetate)
1
to give a colorless sticky liquid 17 (0.300 g, 93.0%). H NMR
(400 MHz, CDCl3): δ 1.38, 1.45 (12 H, s), 3.15 (2 H, m), 3.27 (1 H,
quint), 3.47 (2 H, dd, J = 7.4, 10.7), 3.59 (2 H, dd, J = 10.9, 6.05),
3.71 (4 H, dd, J = 12.0, 5.5), 3.82 (4 H, dd, J = 12.0, 9.5). 13
C
NMR (100 MHz, CDCl3): δ 20.4, 27.0, 49.1, 58.8, 62.3, 63.7,
97.7. EI-HRMS calcd for C15H29NO6 319.1995, found 319.1983.
Tris(1,3-dihydroxy-2-propyl)amine, 9. At 0 °C, trifluoroacetic
acid (1.00 mL) was added to a solution of 17 (1.49 g, 4.67 mmol)
in THF and H2O (25 mL, THF/H2O = 4:1 (v/v)). The resulting
solution was allowed to warm to rt and was left overnight.
Solvent was removed in vacuo. The residue was purified on an
ion exchange column (Amberlite, IR-120, Hþ). The column was
eluted with water first, followed by a solution of aqueous
ammonia (1 M). The solvent was removed to give white solid
tertiary amine 9 (1.10 g, 92.0%), mp >190 °C (dec). Anal. Calcd
for C9H21NO6: C, 45.18; H, 8,85; N, 5.85. Found: C, 45.02; H,
8.82; N, 5.74. 1H NMR (400 MHz, D2O): δ 3.16 (3 H, m), 3.53
(12 H, m). 13C NMR (100 MHz, D2O, methanol, 49.5): δ 57.1,
61.3. 1H NMR (400 MHz, DMSO-d6): δ 2.91 (3H, m), 3.31(12H);
The fact that the 15N chemical shift of 9 did not exhibit the
same sort of anomaly as was found by Wong, et al. for 1
makes sense, because that anomaly arose from the compar-
ison of a “planar” amine to several pyramidal amines, but in
the present case, two “planar” amines (one slightly more
planar than the other) were compared.
(37) (a) Lichter, R. L.; Roberts, J. D. J. Am. Chem. Soc. 1972, 94, 2495–
2500. (b) Levy, G. C.; Holloway, C. E.; Rosanske, R. C.; Hewitt, J. M.;
Bradley, C. H. Org. Magn. Reson. 1976, 8, 643–647. (c) Liepins, E.; Birgele, I.;
Zelcans, G.; Urtane, I.; Lukevics, E. Zh. Obshch. Khim. 1980, 50, 2733–2737.
(38) Baum, J. S.; Shook, D. A.; Davies, H. M. L.; Smith, H. D. Synth.
Commun. 1987, 17, 1709.
4478 J. Org. Chem. Vol. 75, No. 13, 2010

Jie et al.
JOCArticle
4.91 (6H, q); 13C NMR (100 MHz, DMSO-d6): δ 57.1, 60.8. 15
N
(Mo KR) radiation: monoclinic a = 6.8972(4) A, b = 8.2335(5) A,
˚
˚
˚
NMR (40.6 MHz, D2O) δN: 14.23. pKa = 3.08 (T = 25 °C,
0.01 M, titrant: 0.1 N HCl standard solution). E1/2ox = 0.88 V (rt,
0.5 M Na2SO4, in water, Au electrode). A crystal 0.32 mm ꢀ
0.20 mm ꢀ 0.39 mm was selected for X-ray crystallography
c = 11.2422(7) A, R = 92.2850(10)°, β = 102.8470(10)°, γ =
91.2820(10)°; Z = 2; 3019 reflections were collected, 2805 indepen-
dent, -9 e h e 9, -10 e k e 10, -14 e l e 14; full-matrix least-
squares refinement on F2, data-to-parameter ratio =12.4, good-
ness-of-fit =1.068, R1 = 0.0313, wR2 = 0.0842 (all data), R1 =
0.0297, wR2 = 0.0831 (I > 2σ (I)). (b) Reaction of 9 with HCl. A
212.5 mg (0.881 mmol) portion of 9was dissolved in 0.25 mL of 12.1
M HCl (3.0 mmol). Solvent was removed on a rotary evaporator,
and the residue dried further on the vacuum line overnight, afford-
ing 222.0 mg of 9HþCl- (91% yield). This material was pure
by NMR; a portion was recrystallized from methanol, yielding
a colorless crystalline solid, mp 152-152.5 °C. Anal. Calcd for
C9H22NO6Cl: C, 39.20; H, 8.04; N, 5.08. Found: C, 39.21; H 8.20;
N, 5.11.
˚
with 0.71073 A (Mo KR) radiation. Unit cell dimensions a =
˚
˚
˚
10.915(2) A, b=8.9100(18) A, c=23.635(5) A, R = β=γ=90°;
Z=8. Ambient temperature, absorption coefficient=0.114 mm-1
;
21931 reflections were collected, 2858 independent (Rint=0.0239),
-14 e h e 14, -11 e k e 11, -31 e l e 30; full-matrix least-
squares on F2, data-to-parameter ratio =16.8, Goodness-of-fit
1.026, R1=0.0361, wR2=0.1009 (I>2σ (I)), R1=0.0378, wR2=
0.1032 (all data); extinction coefficient=0.0235(19).
Cyclic Voltammetry. The apparatus and procedures used have
been described previously.6 Conditions for 9: 3 mM 0.5 M aq
Na2SO4, 0 °C, Au electrode, 100 mV/s sweep rate, reference
electrode Ag/AgCl satd KCl. At a sweep rate of 20 mV/s, a
reduction wave was in evidence.
Protected Tertiary Amine, 18. At 0 °C, trimethylsilyl chloride
(0.480 mL, 3.78 mmol) was added to a stirred solution of tertiary
amine 9 (100 mg, 0.420 mmol) and imidazole (260 mg, 3.78
mmol) in DMF (10 mL). The mixture was stirred at 0 °C for 1 h
and for 4 h at rt. Water (20 mL) was added to the reaction
solution at 0 °C. The reaction mixture was extracted with ether
(3 ꢀ 20 mL), and the organic layer was washed with brine (2 ꢀ
15 mL), dried with sodium sulfate, and evaporated. The residue
was purified by column chromatography on silica gel using
hexane/ethyl acetate (20:1) as the eluent to give 230 mg (82%) of
compound 18 as a colorless oil. 1H NMR (400 MHz, CDCl3): δ
3.60 (12 H, d), 3.06 (3 H, m), 0.15 (54 H, s); 13C NMR (100 MHz,
CDCl3): δ 62.9, 58.1, -0.3. Anal. Calcd for C27H69NO6Si6: C,
48.23; H, 10.34; N, 2.08. Found: C, 48.51; H 10.56; N, 2.21.
Tris(1,3-dihydroxy-2-propyl)amine Hydrochloride, 9HþCl-.
(a) Unintended formation on reaction of 9 with SiCl4. A solution
of tertiary amine 9 (100 mg, 0.420 mmol) in DMF (3 mL) was
cooled to -5 °C. Silicon tetrachloride (96.5 μL, 0.840 mmol) was
added dropwise. The reaction mixture was warmed gradually to
room temperature, and an initially formed precipitate dissolved.
The reaction was stirred for 4 h at rt (after 1.5 h, the solution became
cloudy and some precipitate formed again). The reaction mixture
was filtered to give 120 mg of a white solid. This was dissolved in
H2O (15 mL) and filtered. The filtrate was concentrated under re-
duced pressure to give a white solid (91.7 mg). This was recrystal-
lized from methanol to give a crystalline solid. 1H NMR (400 MHz,
D2O): δ 4.03 (1 H, m; 2H, m), 3.94 (2H, m); 13C NMR (100 MHz,
D2O): δ 63.8, 57.9. The X-ray crystal structure showed it was
tertiary amine HCl salt 9HþCl-. A crystal 0.278 mm ꢀ 0.165 mm ꢀ
Supporting Information Available: NMR spectra for new
compounds, table of comparisons used to calculate A, CIF files
and ORTEP plots of 9 and 9HþCl-. This material is available
free of charge via the Internet at http://pubs.acs.org.
˚
0.280 mm was selected for X-ray crystallography with 0.71073 A
J. Org. Chem. Vol. 75, No. 13, 2010 4479