Photochemical control of molecular motion associated with pyramidal inversion

Article abstract of DOI:10.1039/c3cc43036g

Rates of pyramidal motion in an aziridine are controlled by reversible trans to cis photoisomerisation of an azobenzene unit attached to the ring nitrogen atom. The dynamics of the inversion process and activation parameters are derived by variable temper

Full text of DOI:10.1039/c3cc43036g

ChemComm
View Article Online
View Journal
COMMUNICATION
Photochemical control of molecular motion associated
with pyramidal inversion†
Cite this: DOI: 10.1039/c3cc43036g
Alexander J. Hough,a Ivan Prokes,a James H. R. Tucker,*b Michael Shipman*a and
Tiffany R. Walsh*ac
Received 24th April 2013,
Accepted 7th June 2013
DOI: 10.1039/c3cc43036g
www.rsc.org/chemcomm
Rates of pyramidal motion in an aziridine are controlled by rever-
sible trans to cis photoisomerisation of an azobenzene unit
attached to the ring nitrogen atom. The dynamics of the inversion
process and activation parameters are derived by variable tem-
perature NMR, and are supported by ab initio calculations.
As part of the continuing interest in the development of
molecular machines and switching devices in chemistry and
biology, light is considered to be an important means through
which external control can be brought to bear on such systems.1
It is well known that reversible photocontrol can be readily
realised through the use of photochromic groups such as
azobenzene.2 There are now many examples of photoswitchable
devices using this approach,2–5 for example in studies on photo-
controllable reactivity and binding processes.3 Another active
area within this field is photocontrollable molecular motion but
despite extensive work on interlocked structures and unidirec-
tional systems,4 there are just a few examples where effective
Scheme 1 Photochemical control of pyramidal inversion rates.
photocontrol over simple dynamic processes such as bond of aziridines and related heterocycles with various coordinating
rotation rates have been achieved.5 Here we demonstrate for groups,8 as well as redox-responsive units,9 can dramatically
the first time, through studies on azobenzene-linked aziridines 1 affect pyramidal inversion rates. Of relevance to this work, it is
and 2 (Scheme 1), how another form of dynamic molecular well known that the rate of pyramidal inversion in simple N-aryl
motion, that of pyramidal atomic inversion, can be reversibly aziridines is very sensitive to the nature of substituents on the
controlled by light.
aromatic ring, with electron withdrawing groups lowering the
A number of factors led to the design and construction of barrier through greater conjugation of the nitrogen lone pair into
photoresponsive aziridines 1 and 2. The use of aziridine based the benzene ring.10 As far as photoresponsive units are concerned,
systems (in this case, 1a # 1b and 2a # 2b) allows accurate the choice of azobenzenes was motivated by knowledge that
measurement of the inversion dynamics using VT-NMR spectro- they have been well studied as elements in light-controlled
scopy.6,7 We have previously demonstrated how functionalisation molecular devices,2 and that methods for the isomerisation
between the cis and trans forms are well established.11 Typi-
cally, trans - cis isomerisation occurs following UV-visible
light irradiation, with the reverse reaction often being conve-
a Department of Chemistry, University of Warwick, Gibbet Hill Road, Coventry,
CV4 7AL, UK. E-mail: m.shipman@warwick.ac.uk; Fax: +44 (0)2476 524112;
niently induced with heat. In this study, we aimed to use
photochemically induced isomerisation to modulate the extent
of conjugation of the nitrogen lone pair into the p-system by
disrupting the planarity of the azobenzene unit. In 1, the trans
azobenzene unit was expected to be planar with significant
conjugation of the trigonal aziridine nitrogen atom in the
transition state (TS) across the entire p-framework, leading to
Tel: +44 (0)2476 523186
b School of Chemistry, University of Birmingham, Edgbaston, Birmingham,
B15 2TT, UK. E-mail: j.tucker@bham.ac.uk; Fax: +44 (0)121 414 4403
c Institute for Frontier Materials, Deakin University, Waurn Ponds, 3216 VIC,
Australia. E-mail: tiffany.walsh@deakin.edu.au; Tel: +61 352 273 116
† Electronic supplementary information (ESI) available: Synthesis and spectro-
scopic data for 1 and 2, procedures for their interconversion, and their activation
barriers (measured and computed). See DOI: 10.1039/c3cc43036g
c
This journal is The Royal Society of Chemistry 2013
Chem. Commun.

View Article Online
Communication
ChemComm
Table 1 Measured and calculated activation parameters for pyramidal inversion
in 1 and 2
a stable, low-lying TS and hence relatively fast rate of pyramidal
motion. Conversely, in cis-2, the extent of this overlap in the TS
would be reduced as a result of a twist12 in the azobenzene
subunit reducing interaction with the distal aromatic ring
and the attached nitro group. This would be expected to
lead to a slower rate of motion. In this way, light-induced
conversion of 1 into 2 should lead to measurable changes in
the rate of pyramidal motion at the aziridine nitrogen atom
(Scheme 1).
Measureda
Aziridine Tc
DG‡coal DG‡ b DH‡ DS‡
MP2 calculationsc
Gas-phase Solventd
trans-1
cis-2
207 38.9
223 42.2
39.2
41.7
34.4 ꢀ16.3 52.0
49.3
52.9
35.3 ꢀ21.4 56.3
a
b
DG‡ and DH‡ in kJ molꢀ1, DS‡ in J Kꢀ1 molꢀ1
.
DG‡ values at 298 K
with estimated errors for 1: ꢂ0.8 kJ molꢀ1 and 2: ꢂ1.1 kJ molꢀ1 (see ESI
c
d
for full details). Using MP2/aug-cc-pVDZ level. Using the implicit
solvent model (toluene).
The synthesis of aziridine 1 was readily achieved in 3 steps
from 4-nitroaniline and b-hydroxyethylaniline (see ESI†).‡ Next,
photochemical conversion of 1 into 2 was investigated. At first,
when 1 in CD2Cl2 (ca. 80 mM) was irradiated using a 125 W
medium pressure Hg lamp for 1 h, using a o355 nm cutoff
filter, no trace of 2 was seen. Fast thermal isomerisation of 2 -
1 could, in part, account for the apparent lack of reactivity of 1.
Indeed, push–pull azobenzenes such as 4-(diethylamino)-40-
nitroazobenzene (3) are known to have unstable cis isomers
which rapidly isomerise back to the trans form.13 Noting that
solvents with a small dielectric constant slow the cis - trans
thermal isomerisation of 3, leading to much longer half-lives of
the cis isomer,13 we repeated the photoisomerisation of 1 in less
polar solvents. Gratifyingly, when 1 in d6-benzene (e = 2.3) was
irradiated for 1 h at room temperature with a o355 nm cutoff
(188–219 K). For each aziridine, the temperature of coalescence
(Tc) was determined and the free energy of activation at
coalescence (DG‡coal) derived. Significantly, 2 (Tc = 223 K)
coalesced at a measurably higher temperature than 1 (Tc
=
207 K), consistent with our prediction of a higher barrier to
N-inversion in 2. In a second approach, dynamic line-shape
simulations of the aziridine A2X2 spin system performed using
WINDNMR15 enabled the rates of exchange over a range of
temperatures to be determined (Fig. 1).8 Eyring plots [ln(k/T) vs.
1/T] were then used to extract the activation parameters for
these aziridines which are presented in Table 1. The measured
difference in inversion barrier between 1 and 2 (DDG‡) is
ca. 3 kJ molꢀ1 [3.3 kJ molꢀ1 (from coalescence) and 2.5 kJ molꢀ1
(from line shape analysis)]. These data confirm our supposition
that inversion occurs faster in 1, with a difference in barrier
that equates to ca. 3-fold difference in the inversion rate at 223 K
(1, k = 5.9 ꢁ 103; 2, 1.9 ꢁ 103 sꢀ1).
1
filter, quantities of 2 were observed by H NMR spectroscopy
(1 : 2; 83 : 17). High levels of conversion of push–pull azo-
benzenes to the cis-isomer are further hampered by the fact that
their UV-vis spectra are broad and poorly resolved, with overlap
of the p - p* and n - p* transitions.11,13 The UV-vis spectrum
of 1 displayed such characteristics with a broad absorption
centred at 416 nm (see ESI†). Several filters were investigated
to increase the production of 2 (see ESI†). For example, the use of
a bandpass filter based on 0.75% w/v iodine in carbon tetra-
chloride,14 which has a modest transmission of B40%, but is
highly selective for lmax = 390 nm, did result in some improve-
ment (1 : 2; 74 : 26). In combination with lower photoirradiation
temperatures to slow the thermal back reaction, sufficient
quantities of 2 could be produced to determine its pyramidal
inversion rate (vide infra). Thus, photoirradiation below 10 1C
(ice-water bath) for 3 h, allowed 2 to be formed as the major
product (1 : 2; 44: 56).§ Similar results were achieved in
d8-toluene (e = 2.4), with the added advantage that NMR spectra
could be recorded down to ꢀ95 1C (see ESI†).
Ab initio calculations were performed to verify the measured
barriers and gain insights into the origin of the differences in
the rates of motion for 1 and 2. To determine the ground state
(GS) and transition state (TS) energies for 1 and 2, the geome-
tries of the minima and TS were optimised at the MP2/6-31G*
level of theory. For 2, a conformational search was conducted to
identify the GS minimum (see ESI† for full details). Single-point
energies (SPE) for the lowest energy structures were then
calculated at the higher MP2/aug-cc-pVDZ level with and
At room temperature, cis-2 isomerised to trans-1 in d8-toluene,
according to first order kinetics, with k = 2 ꢁ 10ꢀ4 sꢀ1 (see ESI†).
This is significantly slower than values reported for classical
push–pull azobenzenes (e.g. 3 (in C6H6): k = 3 ꢁ 10ꢀ2
s
ꢀ1),13
highlighting the effect of ring strain on the ability of the nitrogen
to delocalise into the p-system. From a practical standpoint,
heating 2 in toluene at 60 1C for 10 min resulted in clean
conversion back to 1, providing a simple method to complete
the switching cycle. Several iterations of this cycle (1 - 2 - 1)
were completed without any sign of degradation.¶
To study the molecular motion within 1 and 2, their activa-
tion parameters were determined by two different VT-NMR
methods (Table 1). Firstly, the H NMR spectra were recorded
Fig. 1 Measured (left) and simulated (right) 1H NMR spectrum (d8-toluene,
500 MHz) of aziridine 2 (as a ca. 1 : 1 mixture with 1) over 183–248 K temperature
range. NMR plots for 1 are provided in the ESI.†
1
over a wide range of temperatures above and below coalescence
c
Chem. Commun.
This journal is The Royal Society of Chemistry 2013

View Article Online
ChemComm
Communication
Notes and references
‡ Aziridine 1 has been reported previously by Hallas and Choi who
indicated that it is best prepared by base induced ring closure to form
the aziridine as the final step.16 This approach was adopted herein
although considerable refinement of the synthetic route was required to
obtain 1 in a satisfactory state of purity. Full details are provided in ESI.†
§ Lowering of the photoirradiation temperature to ꢀ20 1C led to no
improvement in the ratio of 1 : 2. Below 10 1C, the cis isomer was
observed to be thermally stable, so that during the VT-NMR experi-
ments (see Fig. 1), the ratio of 1 to 2 did not change.
¶ Aziridine 1 is stable provided it was kept away from sources of acid.
8 In aziridines 1 and 2, a complex AA0BB0 spin system is predicted.17
Fortuitously, the aziridine signals were observed as pairs of singlets at low
temperature, coalescing to one singlet on warming. As such they could be
analysed as A2B2 spin systems, simplifying the line shape analysis.
Fig. 2 Ground state and transition state structures for 1 and 2 optimised at the
1 For reviews, see: (a) J. J. Zhang, Q. Zou and H. Tian, Adv. Mater.,
2013, 25, 378; (b) A. A. Beharry and G. A. Woolley, Chem. Soc. Rev.,
2011, 40, 4422; (c) S. Silvi, M. Venturi and A. Credi, Chem. Commun.,
MP2/aug-cc-pVDZ level.
´
2011, 47, 2483; (d) D. Bleger, Z. Yu and S. Hecht, Chem. Commun.,
2011, 47, 12260; (e) E. R Kay, D. A. Leigh and F. Zerbetto, Angew.
Chem., Int. Ed., 2007, 46, 72; ( f ) G. Mayer and A. Heckel, Angew.
Chem., Int. Ed., 2006, 45, 4900.
2 For a review, see E. Merino and M. Ribagorda, Beilstein J. Org. Chem.,
2012, 8, 1071.
without implicit solvent. The optimised structures are pre-
sented in Fig. 2, with the numerical values given in Table 1.
Although the absolute values derived from these MP2/aug-cc-
pVDZ level calculations using toluene as implicit solvent over-
estimate the barriers for both 1 and 2, they successfully predict
a higher barrier for 2, and compute a difference (DDG‡) of
3.6 kJ molꢀ1 between the cis and trans isomers that agrees
closely with the experimental values (vide supra). Significant
distortion away from planarity as evidenced in the minima and
TS for 2 disrupts conjugation of the aziridine lone pair into the
extended p-system, leading to a higher barrier. This can be
quantified using a natural bond orbital analysis which revealed
not only the greater extent of delocalisation from the nitrogen
into the ring for 1 compared with 2, but also a greater change
in this delocalisation in moving from the GS - TS for 1
(46.3 kJ molꢀ1) compared with 2 (44.5 kJ molꢀ1) (see ESI†).
These studies demonstrate how it is possible to modify the
atomic pyramidal inversion rates in an aziridine using only the
physical inputs of light and heat. Switching 1 - 2 is success-
fully achieved photochemically, with the reverse reaction facili-
tated with simple heating. Careful choice of solvent and
conditions are required to achieve acceptable conversion during
the photoreaction. Experimental and computed differences in
the activation barrier (DDG‡) for the pyramidal motion for 1 and
2 are of the order of 3 kJ molꢀ1, with cis-2 inverting more slowly
due to a twist in the azo system which reduces the extent of
conjugation of the aziridine lone pair into the p-system. The
process proceeds without degradation and can be cycled several
times. As such, it represents the first example of the use of light
to control the rate of pyramidal atomic inversion and it also
illustrates how dynamic molecular motion may be principally
controlled by changes in bond conjugation rather than sterics.5
Future studies will focus on increasing the magnitude of the
rates differences between the two states, and the development of
molecular devices that exploit this phenomena.
3 For recent examples of photochemical control over binding or
reactivity, see: (a) D. Wilson and N. R. Branda, Angew. Chem., Int.
Ed., 2012, 51, 5431; (b) B. M. Neilson and C. W. Bielawski, J. Am.
Chem. Soc., 2012, 134, 12693; (c) J. Manchester, D. M. Bassani,
J. L. H. A. Duprey, L. Giordano, J. S. Vyle, Z. Y. Zhao and J. H. R.
Tucker, J. Am. Chem. Soc., 2012, 134, 10791.
4 (a) M. Baroncini, S. Silvi, M. Venturi and A. Credi, Angew. Chem., Int.
Ed., 2012, 51, 4223; (b) N. Ruangsupapichat, M. M. Pollard,
S. R. Harutyunyan and B. L. Feringa, Nat. Chem., 2011, 3, 53;
(c) G. Haberhauer, Angew. Chem., Int. Ed., 2011, 50, 6415;
(d) V. Serreli, C. F. Lee, E. R. Kay and D. A. Leigh, Nature, 2007,
445, 523; (e) For a recent example of photocontrollable motion in the
form of transport along a molecular track, see: M. J. Barrell,
A. G. Campana, M. von Delius, E. M. Geertsema and D. A. Leigh,
Angew. Chem., Int. Ed., 2011, 50, 285.
5 (a) A. S. Lubbe, N. Ruangsupapichat, G. Caroli and B. L. Feringa,
J. Org. Chem., 2011, 76, 8599; (b) W. T. Sun, Y. T. Huang, G. J. Huang,
H. F. Lu, I. Chao, S. L. Huang, S. J. Huang, Y. C. Lin, J. H. Ho and
J. S. Yang, Chem.–Eur. J., 2010, 16, 11594 and references therein;
(c) M. K. J. ter Wiel, R. A. van Delden, A. Meetsma and B. L. Feringa,
Org. Biomol. Chem., 2005, 3, 4071.
6 W. B. Jennings and D. R. Boyd, in Cyclic Organonitrogen Stereodynamics,
ed. J. B. Lambert and Y. Takechi, VCH, Cambridge, UK, 1992, pp. 105–158.
7 The data analysis is simplified by the fact that 1a = 1b and 2a = 2b.
8 (a) C. T. Hoang, I. Prokes, G. J. Clarkson, M. J. Rowland, J. H. R.
Tucker, M. Shipman and T. R. Walsh, Chem. Commun., 2013,
49, 2509; (b) L. Giordano, C. T. Hoang, M. Shipman, J. H. R.
Tucker and T. R. Walsh, Angew. Chem., Int. Ed., 2011, 50, 741;
(c) M. W. Davies, A. J. Clarke, G. J. Clarkson, M. Shipman and
J. H. R. Tucker, Chem. Commun., 2007, 5078.
9 M. W. Davies, M. Shipman, J. H. R. Tucker and T. R. Walsh, J. Am.
Chem. Soc., 2006, 128, 14260.
10 J. B. Lambert, Top. Stereochem., 1971, 6, 19.
11 H. M. D. Bandara and S. C. Burdette, Chem. Soc. Rev., 2012, 41, 1809.
12 A. Mostad and C. Romming, Acta Chem. Scand., 1971, 25, 3561.
13 (a) P. D. Wildes, J. G. Pacifici, G. Irick, Jr. and D. G. Whitten, J. Am.
Chem. Soc., 1971, 93, 2004; (b) M. Poprawa-Smoluch, J. Baggerman,
H. Zong, H. P. A. Mass, L. De Cola and A. M. Brouwer, J. Phys. Chem.
A, 2006, 110, 11926.
14 S. L. Murov, in Handbook of Photochemistry, ed. M. Dekker, New
York, 1973.
15 H. J. Reich, J. Chem. Educ.: Software, Ser. D, 1996, 3, 2.
16 G. Hallas and J.-H. Choi, Dyes Pigm., 1999, 40, 99.
17 For examples, see (a) R. E. Carter, T. Drakenberg and N. A. Bergman,
J. Am. Chem. Soc., 1975, 97, 6990; (b) H. Nakanishi and
O. Yamamoto, Tetrahedron, 1974, 30, 2115; (c) T. Drakenberg and
J. M. Lehn, J. Chem. Soc., Perkin Trans. 2, 1972, 5, 532.
We acknowledge EPSRC (EP/F021054/1 and EP/F021275/1)
and the University of Warwick for financial support. The
computing facilities of VPAC, Australia, are gratefully acknowl-
edged. TRW thanks VESKI for an Innovation Fellowship.
c
This journal is The Royal Society of Chemistry 2013
Chem. Commun.