The interplay between hydrogen bonding and π-π Stacking interactions in the crystal packing of N1-thyminyl derivatives, and implications for the photo-chemical [2π + 2π]-cycloaddition of thyminyl compounds

Article abstract of DOI:10.1039/c2pp25228g

The solid-state photo-chemical dimerisation of thyminyl derivatives occurs when two thyminyl units are aligned in such a way that the olefinic moieties are separated by a distance of less than 4.2 A. When irradiated with >270 nm UV, the thyminyl olefinic groups undergo [2π + 2π]-cycloaddition to form a dimeric cyclobutane derivative. However, the design and execution of [2π + 2π]-cycloaddition reactions can be challenging due to the requirement to produce molecular crystals with the necessary olefinic alignment. In this investigation, the crystallographic and solid-state photo-chemical reactions of six N1-thyminyl derivatives are studied. Only one derivative, thyminyl propanamide (7), was found to undergo [2π + 2π]-cycloaddition in the crystalline state. As such, quantum chemical methods were employed to study the photo-chemical transition states of the derivatives, as well as the strengths of typical intermolecular interactions that were observed in their crystal structures (such as π-π stacking between the thyminyl rings, Watson and Crick style hydrogen bonding and hydrogen bonding between functional groups of N1 substituents). These results were used to rationalise the solid-state photo-reactivity of more complex bis-thyminyl monomers. The Royal Society of Chemistry and Owner Societies 2012.

Full text of DOI:10.1039/c2pp25228g

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
Photochemical &
Photobiological Sciences
Cite this: Photochem. Photobiol. Sci., 2012, 11, 1938
www.rsc.org/pps
PAPER
The interplay between hydrogen bonding and π–π stacking interactions in the
crystal packing of N1-thyminyl derivatives, and implications for the
photo-chemical [2π + 2π]-cycloaddition of thyminyl compounds†
Priscilla Johnston,a Ekaterina I. Izgorodina*b and Kei Saito*a
Received 29th June 2012, Accepted 16th August 2012
DOI: 10.1039/c2pp25228g
The solid-state photo-chemical dimerisation of thyminyl derivatives occurs when two thyminyl units are
aligned in such a way that the olefinic moieties are separated by a distance of less than 4.2 Å. When
irradiated with >270 nm UV, the thyminyl olefinic groups undergo [2π + 2π]-cycloaddition to form a
dimeric cyclobutane derivative. However, the design and execution of [2π + 2π]-cycloaddition reactions
can be challenging due to the requirement to produce molecular crystals with the necessary olefinic
alignment. In this investigation, the crystallographic and solid-state photo-chemical reactions of six
N1-thyminyl derivatives are studied. Only one derivative, thyminyl propanamide (7), was found to
undergo [2π + 2π]-cycloaddition in the crystalline state. As such, quantum chemical methods were
employed to study the photo-chemical transition states of the derivatives, as well as the strengths of
typical intermolecular interactions that were observed in their crystal structures (such as π–π stacking
between the thyminyl rings, Watson and Crick style hydrogen bonding and hydrogen bonding between
functional groups of N1 substituents). These results were used to rationalise the solid-state photo-
reactivity of more complex bis-thyminyl monomers.
the design and execution of [2π + 2π]-cycloaddition reactions
can be challenging for a number of reasons.
Introduction
Topochemical reactions, such as the solid-state photo-chemical
[2π + 2π]-cycloaddition, represent a common pathway to syn-
thesise cyclic target molecules that possess precisely controlled
stereochemistry. Such structural control over the photo-products
has made possible the synthesis of complex or strained mo-
lecules, such as ladderanes,1–3 cyclophanes,2 hetero-dimers4 and
even some larger molecules, including oligomers and poly-
mers.4,5 Essentially, the [2π + 2π]-cycloaddition involves a
photo-chemical reaction between two olefinic molecules (via
combination of an excited-state and a ground-state species) to
yield a cyclobutane derivative.6 From a synthetic standpoint, the
identification of high-yielding synthetic routes to cyclobutane
compounds is worthy of pursuit, since cyclobutane moieties
occur in a number of natural products and alkaloids.7 However,
Firstly, topochemical control of the reaction dictates that the
pair of reacting olefins should be parallel to one another, and be
separated by a distance of less than 4.2 Å.8 Secondly, the stereo-
selectivity and photo-chemical yields of [2π + 2π]-cycloaddition
reactions tend to be higher when the reactant molecules are ir-
radiated as solids, rather than as solutions.9 Therefore, to obtain
the desired photo-products in useful quantities, one must effec-
tively control the topochemistry of reactant molecules in the
solid-state.
Molecular crystals of the reactants are certainly the ideal
environments for reproducible stereoselective reactions, since
molecules can pack in a rigid and regularly repeating manner.
However, the rational-design of molecules for crystalline-phase
[2π + 2π]-cycloaddition reactions remains challenging due to the
difficulty in predicting the way a given compound will crystal-
lise. Imparting control over the alignment of reactive groups can
thus be an arduous process. Nevertheless, the field of crystal
engineering has afforded a number of examples in which the
reactant molecules have been designed to have a high probability
of packing in the desired crystalline arrangement to favour the
[2π + 2π]-cycloaddition. Exploited interactions have included
non-covalent hydrogen bonding,10 halogen-bonding,11 and aro-
matic stacking;4 metallic coordination atoms have also been
used,12–15 as well as molecular templates,16–18 covalently linked
spacers,19 and inclusion compounds.20–23
aCentre for Green Chemistry, Monash University, Wellington Road,
Clayton, Australia, 3800. E-mail: kei.saito@monash.edu
bSchool of Chemistry, Monash University, Clayton, Victoria, Australia
†Electronic supplementary information (ESI) available: Geometrical
parameters of optimized transition-states; exemplary figures showing the
optimised B3LYP structures of key interactions observed in the crystal
structures; calculated numbers of interactions observed in the crystal
1
structures; H and 13C NMR spectra of derivatives 4, 6, 7 and 9; partial
1H NMR spectra of irradiated 7 (solution-phase) and 9 (solid-phase).
CCDC 888370, 888371, 888372, 888373 and 888374. For ESI and
crystallographic data in CIF or other electronic format see DOI:
10.1039/c2pp25228g
1938 | Photochem. Photobiol. Sci., 2012, 11, 1938–1951
This journal is © The Royal Society of Chemistry and Owner Societies 2012

View Article Online
In contrast to other molecules capable of [2π + 2π]-cyclo-
addition (e.g. derivatives of cinnamic acid,24 coumarins,25 stilba-
zoles,26 etc.), comparatively fewer photo-active crystalline
derivatives have been reported for thymine. Thymine, a nucleic
acid in DNA, is well known to undergo dimerisation by [2π +
2π]-cycloaddition when exposed to UV radiation at wavelengths
close to the thyminyl absorption maxima (>270 nm). The
reverse reaction is also possible upon exposure of the photo-
dimer to shorter UV wavelengths (220–250 nm).27 In biological
systems thymine photo-dimerisation has been implicated as a
causative factor in certain types of cancer,28,29 and much of the
existing literature focuses on understanding thymine dimerisa-
tions in DNA systems.30–32
Since thymine is biologically derived and is also capable of
the retro-[2π + 2π]-cycloaddition, we are interested in its poten-
tial use as a ‘green’ photo-reactive site in new monomer/polymer
systems. Our research has concentrated on the design of bis-thy-
minyl monomers that can undergo reversible photo-polymeris-
ation by a series of [2π + 2π]-cycloaddition reactions. These
types of reversible polymers are important, because they expand
the understanding of [2π + 2π]-cycloaddition and thymine
photo-dimerisation in particular, but also this class of polymers
could be used as recyclable plastics, self-healing polymers, or
optical materials. Although we have already reported on the
reversible topochemical photo-polymerisation of dimethyl 3,3′-
(3,3′-(butane-1,4-diyl)bis(5-methyl-2,4-dioxo-3,4-dihydropyri-
midine-3,1(2H)-diyl))dipropanoate crystals (Scheme 1),33 under-
standing the structural factors necessary to achieve adequate
monomer topochemistry is paramount to the design of new
cyclobutane compounds and reversible polymers. In this paper
we will present the results of a series of experimental and theor-
etical investigations that have enabled us to explain in more
detail why certain monomer features lead to ideal topochemical
arrangements for photo-polymerisation by the [2π + 2π]-
cycloaddition.
Thirdly, quantum chemical calculations at the MP2 level of ab
initio theory were also performed to study the strength of typical
inter-molecular interactions observed in the crystal structures of
N1 thyminyl derivatives, such as π–π stacking between the thy-
minyl rings, Watson and Crick style hydrogen bonding and
hydrogen bonding between functional groups of N1 substituents.
The outcomes of the quantum chemical study highlighted a
series of structural and energetic factors affecting the topochem-
istry of the N1 thyminyl derivatives in the crystalline state and
hence, their photo-reactivity towards photo-dimerisation. These
computational results were then used to rationalise the solid-state
photo-reactivity of bis-thyminyl monomers with the view of
developing a strategy for a smart design of more complex bis-
thyminyl monomers capable of photo-reactions in the crystalline
state.
Experimental
General
All chemicals were used as supplied by Sigma-Aldrich, and
1
were of the highest available purity. H- and 13C-nuclear mag-
netic resonance experiments were conducted on a Bruker Avance
400 Spectrometer. All spectra were recorded at 25 °C, and refer-
enced to the residual solvent signal. Infrared spectra were
recorded using either a Perkin Elmer 2000 FT-IR spectro-
photometer, in absorbance mode using KBr as the background
reference; or, a Bruker Equinox 55 in ATR mode with diamond
as the background reference.
X-ray diffraction
We adopt a “bottom-up” approach to the understanding of
photo-reactive bis-thyminyl monomers. Firstly, we investigate
the solid-state photo-reactions of six simple N1 thyminyl deriva-
tives: thyminyl acetic acid (2), thyminyl methyl acetate (3), thy-
minyl acetamide (4), thyminyl methyl propanoate (5), thyminyl
propanoic acid (6) and thyminyl propanamide (7). Among the
six monomers studied, only thyminyl propanamide 7 was suc-
cessfully dimerised with UV radiation in the crystalline state.
Secondly, the photo-reactivity of the monomers was studied
through the concerted mechanism of the photo-induced [2π +
2π] cycloaddition using standard transition state theory34
coupled with a correlated wavefunction-based method, MP2.
Due to the small or weakly diffracting crystals obtained for 3, 5,
6 and 9, the reported structural analyses were performed on the
MX1 micro-crystallography beam-line at the Australian Synchro-
tron, Clayton, Victoria. The end station comprised a φ goniostat
with a Quantum 210r area detector. Data were collected using
the Blue Ice GUI35 and processed using the XDS software.36
Due to hardware constraints (fixed detector angle, minimum
detector distance) the maximum obtainable resolution at the
detector edge was approximately 0.81 Å.
A single crystal of derivative 7 was coated in viscous oil and
mounted on a glass fibre. Data (2θmax 50–55°) were collected at
173(2) K using an Enraf Nonius KAPPA CCD and MoKα
(λ 0.71073 Å) radiation. After integration and scaling, datasets
were merged (Rint as quoted). Data were corrected for absorption
using SORTAV.37
All the structures were solved and refined using the programs
SHELXS-9738 and SHELXL-97,39 respectively. The program
X-Seed40 was used as an interface to the SHELX programs, and
for preparation of the figures. Plausible positions of hydrogen
atoms in water molecules, carboxylic groups and amides, were
located in the difference Fourier map, and were refined such that
O–H distances were restrained to reasonable values
(0.88–0.98 Å); all other hydrogen atoms were placed in calcu-
lated positions using a standard riding model.
Scheme 1 Reversible photo-polymerisation of 3,3′-(3,3′-(butane-1,4-
diyl)bis(5-methyl-2,4-dioxo-3,4-dihydropyrimidine-3,1(2H)-diyl))dipro-
panoate (8) by the [2π + 2π]-cycloaddition reaction of thyminyl units.33
This journal is © The Royal Society of Chemistry and Owner Societies 2012
Photochem. Photobiol. Sci., 2012, 11, 1938–1951 | 1939

View Article Online
Synthesis of the N1-thyminyl derivatives (2–7)
P21/c (no. 14), a = 8.9270(18), b = 14.369(3), c = 7.7640(16) Å,
β = 90.60(3)°, V = 995.8(4) Å3, Z = 4, Dc = 1.415 g cm−3, F000
= 448, goniostat with quantum 210r detector, synchrotron radi-
ation, λ = 0.71253 Å, T = 100(2) K, 2θmax = 50.0°, 12566 reflec-
tions collected, 1702 unique (Rint = 0.2761). Final GooF =
1.059, R1 = 0.0841, wR2 = 0.2074, R indices based on 1572
reflections with I > 2σ(I) (refinement on F2), 143 parameters,
0 restraints. Lp and absorption corrections applied, μ =
0.113 mm−1. As only weakly diffracting crystals were obtained
for 5, a large proportion of essentially “unobserved” reflections
were used in the refinement which probably leads to the elevated
value for Rint (0.276). Nevertheless, a satisfactory refinement
was obtained (R1 = 0.0841).
2-(5-Methyl-2,4-dioxo-3,4-dihydropyrimidin-1(2H)-yl)acetic
acid (2). Acetic acid, 2, was synthesised according to a reported
method.41 A crystal structure has also been reported for this
compound.42
Methyl 2-(5-methyl-2,4-dioxo-3,4-dihydropyrimidin-1(2H)-yl)-
acetate (3). Methyl acetate 3 was prepared using a reported
method.43 Single crystals of 3 were obtained upon recrystallisa-
tion from hot EtOH.
Crystal data (CCDC 888370): C8H10N2O4, M = 198.18, col-
ourless needle, 0.3 × 0.01 × 0.01 mm3, monoclinic, space group
P21/n (no. 14), a = 5.0700(10), b = 22.070(4), c = 8.3100(17) Å,
β = 100.85(3)°, V = 913.2(3) Å3, Z = 4, Dc = 1.441 g cm−3, F000
= 416, goniostat with quantum 210r detector, synchrotron radi-
ation, λ = 0.71068 Å, T = 100(2) K, 2θmax = 50.0°, 5903 reflec-
tions collected, 1526 unique (Rint = 0.0522). Final GooF =
1.050, R1 = 0.0469, wR2 = 0.1204, R indices based on 1349
reflections with I > 2σ(I) (refinement on F2), 133 parameters,
0 restraints. Lp and absorption corrections applied, μ =
0.117 mm−1. The measured completeness is low (0.956) due to
the hardware limitations of the synchrotron beamline (i.e. fixed
detector angle, minimum detector distance). However, the
amount of observed data is close to 100% and yields a more
than satisfactory refinement (R1 = 0.0469). A short intermolecu-
lar contact was also observed between O7⋯C11 (2.96 Å), which
was consistent with the Type I Oδ−⋯δ+CvO (carbonyl–carbo-
nyl) perpendicular interaction motif described by Allen et al.44
3-(5-Methyl-2,4-dioxo-3,4-dihydropyrimidin-1(2H)-yl)propanoic
acid (6). The propanoic acid derivative 6, was prepared by the
base hydrolysis of the methyl ester 547 (1.06 g, 5.0 mmol) in
aqueous NaOH (1 M, 20 mL). The reaction mixture was heated
at reflux for 3 h, and the solution was cooled and acidified to pH
2.0 with 10% HCl. The title compound, 6, crystallised from the
cooled solution (2 °C), and was washed repeatedly with H2O.
The NMR spectral data were consistent with literature data.
Yield: 0.90
g
(91%). M.p. 181.7–182.9 °C (lit.48
172–174 °C). MS (ESI+): Calcd for C8H10N2O4: m/z 198.1;
Found: m/z 199.2 (M + H+), 221.2 (M + Na+). 1H NMR
(400 MHz, D6-DMSO): δH 1.73 (d, J = 1.2 Hz, 3H, C5-CH3),
2.59 (t, J = 7.2 Hz, 2H, CH2CO), 3.81 (t, J = 7.2 Hz, 2H, N1-
CH2), 7.49 (d, J = 1.2 Hz, 1H, H6), 11.21 (s, 1H, NH). 13C
NMR (100 MHz, D6-DMSO): δC 11.97 (C5-CH3), 32.90
(CH2CO), 43.94 (N1-CH2), 108.17 (C5), 141.88 (C6), 150.80
(C2), 164.35 (C4), 172.33 (COOH). Selected IR bands (KBr,
cm−1): 3152s, 3090w, 3030s, 2978w, 2940w, 2812m, 1728s,
1696s, 1676s, 1474m, 1458m, 1436m, 1414m, 1402m, 1378m,
1356m, 1252m, 1226m, 1194m, 1160m, 1126m, 1054m.
2-(5-Methyl-2,4-dioxo-3,4-dihydropyrimidin-1(2H)-yl)acetamide
(4). Acetamide 4 was obtained by aminolysis of methyl acetate
2. In this procedure, 2 (0.95 g, 4.8 mmol) was added to 15
equivalents of a concentrated aqueous ammonia solution. The
flask was capped and the contents were shaken until the solid
had completely dissolved. The title compound crystallised over-
night from the resulting solution. A pure crystalline sample of 4
was obtained upon recrystallisation from H2O. The NMR spec-
tral data were consistent with literature data.
Crystal data (CCDC 888372): C8H10N2O4, M = 198.18, col-
ourless prism, 0.12 × 0.07 × 0.03 mm3, monoclinic, space group
P21/n (no. 14), a = 4.9130(10), b = 20.405(4), c = 8.5920(17) Å,
β = 101.90(3)°, V = 842.8(3) Å3, Z = 4, Dc = 1.562 g cm−3, F000
= 416, goniostat with quantum 210r detector, synchrotron radi-
ation, λ = 0.71072 Å, T = 100(2) K, 2θmax = 50.0°, 8717 reflec-
tions collected, 1336 unique (Rint = 0.1273). Final GooF =
1.075, R1 = 0.0577, wR2 = 0.1552, R indices based on 1245
reflections with I > 2σ(I) (refinement on F2), 136 parameters,
0 restraints. Lp and absorption corrections applied, μ =
0.127 mm−1. The measured completeness is low (0.903) due to
the hardware limitations of the synchrotron beamline (i.e. fixed
detector angle, minimum detector distance). However, the
amount of observed data is close to 100% and yields a more
than satisfactory refinement (R1 = 0.0577). As only weakly dif-
fracting crystals were obtained, a large proportion of essentially
“unobserved” reflections were used in the refinement which
probably lead to the elevated value for Rint (0.127).
Yield: 0.72 g (82%). M.p. 297.5–300.2 °C (dec) (lit.45
299–302 °C, dec). MS (ESI−): Calcd for C7H9N3O3: m/z 183.1;
1
Found: m/z 182.2 (M − H). H NMR (400 MHz, D6-DMSO):
δH 1.75 (d, J = 1.2 Hz, 3H, C5-CH3), 4.23 (s, 2H, N1-CH2),
7.18 (s, 1H, amide NH), 7.41 (d, J = 1.2 Hz, 1H, H6), 7.56 (s,
1H, amide NH). 13C NMR (100 MHz, D6-DMSO): δC 11.90
(C5-CH3), 49.15 (N1-CH2), 107.87 (C5), 142.42 (C6), 151.05
(C2), 164.50 (C4), 168.83 (CONH2). Selected IR bands (KBr,
cm−1): 3410bs, 3296bs, 3156bs, 3076w, 3026bs, 2958w, 2934w,
2906w, 2860w, 2822w, 2774m, 1678bs, 1476s, 1418m, 1408m,
1396m, 1354m, 1230s, 1150m.
A crystal structure has already been reported for this
compound.46
Methyl 3-(5-methyl-2,4-dioxo-3,4-dihydropyrimidin-1(2H)-yl)-
propanoate (5). The title compound, 5, was synthesised using a
reported Michael addition reaction between methyl acrylate and
thymine (1).47 A crystalline sample was prepared upon recrystal-
lisation of 5 from hot EtOH.
3-(5-Methyl-2,4-dioxo-3,4-dihydropyrimidin-1(2H)-yl)propan-
amide (7). Thymine propanamide (7) was prepared by aminoly-
sis of methyl propanoate 547 (1 g, 4.7 mmol) using a similar
procedure to that used for the synthesis of acetamide 4. After
recrystallisation from H2O, pure 7 was isolated in 52% yield.
Crystal data (CCDC 888371): C9H12N2O4, M = 212.21, col-
ourless plate, 0.40 × 0.10 × 0.02 mm3, monoclinic, space group
1940 | Photochem. Photobiol. Sci., 2012, 11, 1938–1951
This journal is © The Royal Society of Chemistry and Owner Societies 2012

View Article Online
Yield: 0.48 g (52%). M.p. 233.9–235.8 °C. MS (ESI): Calcd
for C8H11N3O3: m/z 197.1; Found: m/z 220.1 (M + Na). 1H
NMR (400 MHz, D6-DMSO): δH 1.72 (d, J = 0.8 Hz, 3H, C5-
CH3), 2.42 (t, J = 6.8 Hz, 2H, CH2–CvO), 3.80 (t, J = 6.8 Hz,
2H, N1-CH2), 6.89 (s, 1H, amide NH), 7.40 (s, 1H, amide NH),
7.43 (d, J = 1.2 Hz, 1H, C6H), 11.19 (s, 1H, N3H). 13C NMR
(100 MHz, D6-DMSO): δC 11.96 (C5-CH3), 33.91 (CH2CO),
44.39 (N1-CH2), 107.92 (C5), 142.01 (C6), 150.75 (C2), 164.35
(C4), 171.69 (COOMe). IR (KBr, cm−1): 3410m, 3364m,
3232m, 3188m, 3078m, 3048w, 2964w, 2932w, 2804w, 1686s,
1672s, 1658s, 1618s, 1472m, 1464m, 1448w, 1434w, 1386m,
1362m, 1242m, 1218m.
2θmax = 50.0°, 7070 reflections collected, 1854 unique (Rint =
0.0760). Final GooF = 1.083, R1 = 0.0500, wR2 = 0.1326, R
indices based on 1685 reflections with I > 2σ(I) (refinement on
F2), 167 parameters, 0 restraints. Lp and absorption corrections
applied, μ = 0.114 mm−1. The water molecule was refined such
that O–H distances were restrained to reasonable values
(0.88–0.98 Å). The measured completeness is low (0.924) due to
the hardware limitations of the synchrotron beamline (i.e. fixed
detector angle, minimum detector distance). However, the
amount of observed data is close to 100% and yields a more
than satisfactory refinement (R1 = 0.0500).
Crystal data (CCDC 888373): A crystal structure has been
reported for this compound,49 but the data presented below gave
an improved refinement. C8H11N3O3, M = 197.20, colourless
needle, 0.16 × 0.12 × 0.10 mm3, monoclinic, space group P21/c
(no. 14), a = 7.3571(15), b = 15.243(3), c = 15.854(3) Å, β =
Photo-reactivity studies
All of the [2 + 2]-cycloaddition reactions were performed using
an Ultraviolet Products CL1000M UV-crosslinker lamp that pro-
duced mid-range UV-wavelengths centred at 302 nm.
90.18(3)°, V = 1777.9(6) Å3, Z = 8, Dc = 1.473 g cm−3, F000
=
832, Nonius Kappa CCD, MoKα radiation, λ = 0.71073 Å, T =
173(2) K, 2θmax = 50.0°, 10 151 reflections collected, 2947
unique (Rint = 0.0408). Final GooF = 1.060, R1 = 0.0370, wR2 =
0.0814, R indices based on 2350 reflections with I > 2σ(I)
(refinement on F2), 279 parameters, 0 restraints. Lp and absorp-
tion corrections applied, μ = 0.115 mm−1. Although the
measured completeness is 0.941, the amount of observed data is
close to 100% and yields a more than satisfactory refinement
(R1 = 0.0370).
Irradiation of solutions of 2–7. An aqueous solution of com-
pound 7 (5 mL, 4.0–5.0 × 10−2 M) was transferred to a sealable
pyrex tube (10 mL), purged with N2, and then irradiated with
302 nm UV (0.35 kJ cm−2). The solution was then freeze-dried
and the resulting crude solids were subjected to H NMR analy-
sis in order to determine the extent of photo-dimerisation.
1
Irradiation of crystalline samples. Crystalline samples
(approximately 100 mg) were irradiated as thin layers of crystals
in uncovered petri dishes. The relevant irradiation doses are
specified but generally ranged between 0–0.53 kJ cm−2. All
photo-reactions were monitored by 1H NMR.
Synthesis of the N3-butyl bridged bis-thyminyl derivatives
(8–10)
The syntheses and crystal structures of bis-methyl propanoate (8)
and bis-propanamide (10) were described previously.7
Theoretical procedures
Standard methods for the orbital approximation were used
within GAUSSIAN 09.50 All six thyminyl derivatives (thyminyl
methyl propanoate, thyminyl propanoic acid, thyminyl methyl
acetate, thyminyl acetic acid, thyminyl propanamide and thymi-
nyl acetamide) were considered in the computational study. The
geometry optimisations of the ground state of these thyminyl
monomers and their dimers were performed at the B3LYP/
6-31G(d) level of theory.51–54 Improved electronic energies were
obtained at the MP2 level of theory55 in conjunction with the cc-
pVTZ basis set.56 The geometries of the monomers were confor-
mationally screened to ensure the lowest energy structures. The
dimers in this work were defined as two monomers non-cova-
lently interacting through either π–π stacking or hydrogen
bonding. The binding energies of dimers were calculated as the
difference between the energy of the dimer and those of two
monomers. All binding energies were corrected for zero-point
vibration energy taken out of the B3LYP frequency calculations.
Due to the similar size of the thyminyl derivatives studied the
basis set superposition error (BSSE) was assumed to equally
affect the binding energies of the dimers and therefore, time-con-
suming BSSE calculations were excluded from this study. Stan-
dard transition state theory34 was used to locate transition states
following the concerted mechanism of [2π + 2π] cycloaddition
at the B3LYP/6-31G(d) level of theory for four thyminyl deriva-
tives: methyl acetate 3, propanoic acid 6, methyl propanoate 5
and propanamide 7. The monomers in their lowest energy
3,3′-(3,3′-(Butane-1,4-diyl)bis(5-methyl-2,4-dioxo-3,4-dihydro-
pyrimidine-3,1(2H)-diyl))dipropanoic acid (9). The bis-methyl
propanoate monomer 8 (1.3 g, 2.8 mmol),7 was heated to reflux
in aqueous KOH (3 M, 30 mL) for 4 h or until the reaction
mixture became a clear solution. Whilst still hot, the reaction
mixture was rapidly acidified with 10% aqueous HCl to yield
single crystals of the title compound. The crystals were filtered
and washed repeatedly with H2O.
Yield: 0.97 g, 80%. M.p. 236.8–239.2 °C. MS (ESI): Calcd
for C20H26N4O8: m/z 450.18; Found: m/z 473.0 (M + Na,
1
100%). H NMR: (400 MHz, D6-DMSO) δ 1.47 (br, s (p), 4H,
core CH2), 1.78 (s, 3H, C5-CH3), 2.61 (t, J = 6.8 Hz, 4H,
CH2CO), 3.78 (br, s (t), 4H, N3-CH2), 3.87 (t, J = 6.8 Hz, 4H,
2 × N1-CH2), 7.56 (s, 1H, C6H). 13C NMR (100 MHz, CDCl3):
δC 12.54 (C5-CH3), 24.60 (alk. CH2), 32.70 (CH2CvO), 40.10
(N3-CH2), 44.95 (N1-CH2), 107.27 (C5), 140.47 (C6), 150.64
(C2), 163.03 (C4), 172.22 (COOH). IR (ATR, cm−1): 3498 m,
3071 m, 2930 m, 1714 m, 1691 m, 1621 s, 1434 m, 1383 m,
1357 m, 1205 m.
Crystal data (CCDC 888374): C20H30N4O10, M = 486.48, col-
3
ˉ
ourless prism, 0.07 × 0.03 × 0.01 mm , triclinic, space group P1
(no. 2), a = 7.5950(15), b = 8.9000(18), c = 9.4540(19) Å, α =
96.93(3), β = 111.13(3), γ = 101.53(3)°, V = 570.9(2) Å3, Z = 1,
Dc = 1.415 g cm−3, F000 = 258, goniostat with quantum 210r
detector, synchrotron radiation, λ = 0.71069 Å, T = 100(2) K,
This journal is © The Royal Society of Chemistry and Owner Societies 2012
Photochem. Photobiol. Sci., 2012, 11, 1938–1951 | 1941

View Article Online
structures were taken for the transition state search. Four possible
configurations of the transition state, such as trans–syn (TS),
trans–anti (TA), cis–syn (CS) and cis–anti (CA) for each
monomer were considered. Improved reaction barriers were cal-
culated at the MP2/cc-pVTZ level of theory. The calculated reac-
tion barriers were also corrected for zero-point vibration energy
from B3LYP.
literature reports or were in agreement with the proposed
structures.
Solid-state dimerisation and the influence of crystal packing
Crystalline samples of the synthesised N1 derivatives were
obtained from either ethanol (esters 3 and 5) or water (acid and
amide derivatives), and these samples were used for both the
X-ray crystallographic and photo-chemical studies. The unit cells
collected for crystals of 2 and 4 were consistent with previously
reported crystal structures, and in these cases, references are
made to the literature structures.42,46 In order to examine the
solid-state photo-reactivity of the thyminyl derivatives, the crys-
talline samples were each irradiated with 302 nm UV. However,
only amide 7 was found to undergo [2π + 2π]-cycloaddition to
form the corresponding cyclobutane dimer (Fig. 1).
Referring to Table 1, the crystal structures of compounds 2–6
each possessed olefinic separation distances of more than 4.2 Å,
which accounted for their photo-stability. Only crystals of propa-
namide 7, whose olefinic moieties were separated by 3.67 and
3.76 Å, underwent [2π + 2π]-cycloaddition to form the corres-
ponding cyclobutane dimers in a total yield of 56%.
Results and discussion
To perform high yielding photo-chemical syntheses of thyminyl
cyclobutane derivatives, the reactive olefins must be appropri-
ately positioned in the crystal lattice, such that the olefinic
groups align with one another and are separated by a distance of
less than 4.2 Å.8 In the literature to date, systematic studies con-
cerning the photo-reactivity and crystal-packing behaviour of
crystalline N1-thyminyl derivatives are limited to investigations
involving 1-alkylthymines (from C5H11 to C10H21)57 and -alkyl-
propanoates (from C9H19 to C16H33),58 which possess long and
weakly interacting alkyl chains. Thus, it was of interest to study
the photo-reactivity and crystal packing behaviour of shorter-
chain thyminyl derivatives, bearing either a CH2CO or C2H4CO
N1-spacer. In the previous studies, the thyminyl derivatives were
decorated with only weakly interacting alkyl chains.57,58 It is
therefore important to examine how varying the N1 functionality
between esters (3 and 5), carboxylic acids (2 and 6), and amides
(4 and 7) can influence the molecular crystal packing through
hydrogen bonding and π–π stacking interactions between the
olefinic groups, and therefore the photo-reactivity of the crystal-
line derivatives.
The crystal structure of 7 in Fig. 2 reveals a trans–syn (TS)
type alignment between pairs of 7 prior to the irradiation. There-
fore, according to the accepted topochemical arguments of
Schmidt,8 the main photo-dimer isomer should also possess the
TS stereochemistry. Indeed, when the photo-dimerisation of 7
was followed throughout the irradiation (0–0.53 kJ cm−2) using
1H NMR, the TS isomer was the major cyclobutane dimer
formed (obtained in 44% yield) (Figs. 1 and 2). Upon examin-
1
ation of the H NMR spectrum of irradiated 7 (Fig. 1), some
small amounts of the cis–anti (CA), cis–syn (CS) and trans–anti
Synthesis of the N1-thyminyl derivatives (2–7)
(TA) dimers were also detected; however, these species were
Thyminyl derivatives (2–7) were synthesised according to
Scheme 2, in moderate to good yields. Briefly, thymine (1) was
alkylated at the N1 position using bromoacetic acid to form the
acetic acid derivative, 2. The corresponding methyl ester 3 was
obtained by H2SO4 catalysed esterification of acid 2 in methanol.
Ester 3 was subjected to aminolysis using NH3 to obtain amide
4. Propanoate 5 was synthesised by Michael addition of methyl
acrylate to thymine (1). The corresponding carboxylic acid (6)
and amide (7) derivatives were subsequently formed by the base
hydrolysis or aminolysis of ester 5, respectively. The spectro-
scopic analyses of each derivative were consistent with previous
Fig. 1 Partial 1H NMR spectrum of the irradiated sample of crystalline
7 showing peaks corresponding to the C5–CH3 protons of monomer (M)
and cyclobutane dimers (CA, CS, TA and the major product, TS).
Scheme 2 The synthesis of the N1-thyminyl derivatives.
1942 | Photochem. Photobiol. Sci., 2012, 11, 1938–1951
This journal is © The Royal Society of Chemistry and Owner Societies 2012

View Article Online
Table 1 Measured distances (given in Å) of the observed interactions in the crystal structures of 2–7
Crystal system/ Olefinic separation
No. space group
distance (Å)
Type of π–π stacking
WC HB (Å) Inter-chain HB (Å)
Chain–ring HB (Å)
2
3
4
5
6
7
Monoclinic
P21/n 42
Monoclinic
P21/n
5.04
CS, slipped and displaced 2.05
CS, slipped and displaced 1.95
NA
COOH⋯OvC4 1.88
C2vOδ− δ+CvO 2.96
5.07
NA
⋯
Monoclinic
5.21
TA, displaced sandwich
TA, sandwich
1.93a
1.96
CONH2⋯OvCNH2 2.14
CONH2⋯OvC′2 2.14
NA
P21/c 46
Monoclinic
P21/c
Monoclinic
P21/n
Monoclinic
P21/c
4.39
NA
NA
4.91
CS, parallel displacement 1.97
TS, criss-cross NA
COOH⋯OvC′2 1.78
3.67, 3.76
CONH2⋯OvCNH2 2.06, 1.90 CONH2⋯OvC′2 2.02
a Anti-Watson and Crick hydrogen bonding was observed in the crystal structure of acetamide 4.
Fig. 2 (a) Follows the photo-dimerisation of 7 (crystalline-state) to the
trans–syn (TS) cyclobutane derivative with 302 nm UV irradiation
(0.53 kJ cm−2). (b) The crystal structure of propanamide 7 (H atoms not
shown) shows the TS alignment of the closest thyminyl pair. The
carbons of the reactive olefinic groups are displayed as spheres. (c, d)
The optical micrographs of the propanamide crystals before (c) and after
(d) irradiation (magnification ×400) are also compared. The irradiated
propanamide crystals (d) possess fracture lines.
Fig. 3 Crystal structure examples of the: Watson and Crick hydrogen-
bond (WC HB) in propanoic acid 6; inter-chain HB in propanamide 7;
chain–ring HB in propanoic acid 6; and TA π–π stacking in
propanoate 5.
intermolecular interactions were observed in the crystal struc-
tures of the six monomers: (1) π–π stacking interaction between
the thyminyl rings, (2) Watson and Crick style hydrogen
bonding (WC HB) between the thyminyl rings as shown in
Fig. 3, (3) hydrogen bonding between terminal functional groups
on the N1 substituents (Inter-Chain HB) and (4) hydrogen
bonding between the carbonyl group on the thyminyl ring and
functional group on the N1 substituent (chain–ring HB).
only formed as minor photo-products. The formation of the
other dimer isomers could result from either the destruction of
the crystal lattice during photo-reaction or minor faults in the
crystal packing of 7. The accumulation of strain in the lattice as
a result of small reorientations of the reacting molecules during
photo-dimerisation is evidenced by the micrograph in Fig. 2d,
which reveals severe fractures in the irradiated crystals of 7.
To explain the photo-dimerisation success in the case of the
propanamide derivative (7), a thorough analysis of the crystal
structures of the six monomers was performed with the view of
identifying the factors contributing to this result.
Analysis of Table 1 reveals that in most of the crystal struc-
tures, the monomers formed strong Watson and Crick style
hydrogen
bonds
between
the
N3H⋯OvC4
and
C4vO⋯HN3 groups of two adjacent thyminyl rings (Fig. 3).
An anti-Watson and Crick hydrogen bond between
N3H⋯OvC2 and C2vO⋯HN3 (Fig. 4c) was observed in the
crystal structure of acetamide (4), while the propanamide did not
form this type of intermolecular interaction. The lengths of the
hydrogen bonds varied slightly between the N1 derivatives, but
collectively ranged between 1.93 and 2.05 Å. Compared to the
adenine–thymine base pair in DNA, which possesses hydrogen-
X-ray crystallography and structural comparison of the N1
derivatives
Examination of the crystal structures of 2–7 revealed several
common features that are summarised in Table 1. Four typical
This journal is © The Royal Society of Chemistry and Owner Societies 2012
Photochem. Photobiol. Sci., 2012, 11, 1938–1951 | 1943

View Article Online
C4 carbonyl oxygen was also involved in a partial Watson and
Crick hydrogen bond to the N3 hydrogen of another thyminyl
ring.
Another interesting chain–ring interaction was observed in
the crystal structure of 3. In this case, the C2 carbonyl
oxygen and the ester carbonyl carbon of another molecule
were closely associated (d = 2.61 Å). This CvOδ−⋯δ+CvO
contact is a special type of interaction, which is based on weak
electrostatic attractions between the polarised carbonyls, specifi-
cally the partially negative oxygen and partially positive
carbon.44
From the crystal structures, it appeared that the Watson and
Crick pairing stabilised the position of ring stacks, while the
inter-chain and chain–ring interactions effected the arrangement
of proximity related thyminyl ring pairs, thereby influencing the
type of π–π interactions observed in the crystal structures. Co-
operative effects between hydrogen-bonding environment and
π–π stacking have also been proposed for base-stacking in DNA
systems.61 This idea is further strengthened when one considers
the crystal structure of propanoate 5, where the chain–chain and
chain–ring interactions are absent, but Watson and Crick pairing
is still observed. In methyl propanoate 5, the proximity related
thyminyl pairs are generated by TA sandwich stacking, whilst
the position and separation between the stacked pairs is dictated
by the length of the Watson and Crick hydrogen-bonds.
Although propanoate 5 crystallises in the preferable TA thyminyl
ring arrangement, the olefinic groups are separated by 4.4 Å,
which is most likely too long for the [2π + 2π]-cycloaddition to
occur.
Fig. 4 (a) Watson and Crick hydrogen bonding between a thymine–
adenine pair, (b) Watson and Crick style hydrogen bonding observed in
the N1 thyminyl derivatives and (c) anti-Watson and Crick hydrogen
bond observed between molecules of thyminyl acetamide 4.
Fig. 5 Variations of the observed ring arrangements in crystals of 2–7:
(a) Slipped and displaced CS pair in 1 and 2, (b) sandwiching in the TA
pairs of 4 and 5, and (c) parallel displacement in the CS pair in 6, (d)
criss-cross observed in the TS pair of 7.
CS displacements that were stabilised by chain–ring inter-
actions were the most commonly observed packing arrangements
(Fig. 5a,c). In methyl acetate 3, the carbonyl–carbonyl
(C2vOδ−⋯δ+CvO) interaction led to the slipped and displaced
CS packing. In propanoic acid 6, COOH⋯OvC2 and
C2vO⋯HOOC hydrogen bonds produced displaced CS pairs;
bond distances of 2.82 and 2.94 Å,59 the Watson and Crick base
pairing between thyminyl rings in the crystal structures of the
N1 derivatives is much stronger. For N1 octyl thymine, the dis-
tance of the Watson and Crick hydrogen-bonds ranged between
2.77–2.88 Å.60
In the amide derivatives, 4 and 7, inter-chain hydrogen
bonding between an amide hydrogen atom and the amide carbo-
nyl oxygen of a neighbouring molecule (NH2CvO⋯HN(H)
CvO) appeared to be slightly stronger in the propanamide
crystal as the measured hydrogen-bond distance of 1.94 Å was
shorter than that of acetamide 4 (2.14 Å).46 The longer bond dis-
tance in acetamide 4 may have been due to competition with the
π–π stacking interactions between thyminyl rings, since in aceta-
mide the TA-type sandwich arrangement (Fig. 5b) of the thymi-
nyl rings was slightly displaced as a result of the inter-chain
hydrogen bond.
In each of the N1 derivatives that possessed a hydrogen-bond
donor atom (i.e. COOH and CONH2) in the N1 chain, chain–
ring hydrogen-bonding interactions were also observed. In
amides 4 and 7, the hydrogen bond occurred between an amide
hydrogen atom and the C2 carbonyl oxygen of the thyminyl ring
in a neighbouring molecule. Similarly in propanoic acid 6
the hydrogen bond occurred between the COOH hydrogen and
the C2 carbonyl oxygen. However, in acid 2, the chain–ring
interaction occurred between the COOH hydrogen and neigh-
bouring C4 carbonyl oxygen. The hydrogen bond to the
C4 carbonyl oxygen in acid 4 was unusual considering that the
while
a
combination of chain–ring hydrogen bonding
(COOH⋯OvC4, C4vO⋯HOOC) and WC hydrogen bonding
contributed to the slipped and displaced CS ring packing in
the crystal structure of acetic acid 2. It is also interesting to note
that the photo-reactive propanamide 7 crystals included a criss-
cross orientation of the olefinic groups (Fig. 5d). This ring orien-
tation was stabilised by extensive inter-chain and chain–ring
hydrogen bonding, rather than a Watson and Crick hydrogen
bond.
Of the six monomers only propanamide 7, which displayed
π–π stacking between the TS criss-crossed olefins, underwent
photo-dimerisation. The remaining derivatives, 2–6, were photo-
stable. These observations prompted us to computationally
investigate two aspects of the photo-chemical reactions. Firstly,
it was of fundamental interest to determine whether the
reaction barriers for the photo-induced [2π + 2π] cycloadditions
vary when the N1 functionality is changed, or when
different π–π stacking interactions (i.e. TA, TS, CA and CS)
are adopted. Secondly, it was useful to examine the
energetic differences between the four types of intermolecular
interactions identified in the crystal structures described above,
and to also investigate the roles that these interactions ultimately
play in the topochemistry of the thyminyl derivatives in the crys-
talline state.
1944 | Photochem. Photobiol. Sci., 2012, 11, 1938–1951
This journal is © The Royal Society of Chemistry and Owner Societies 2012

View Article Online
Table 2 MP2 reaction barriers for the concerted mechanism of photo-
induced [2 + 2] cycloaddition
Substituent effects on reaction barriers for the concerted
mechanism
Monomer
Type of cycloaddition
ΔH#, kJ mol−1
ω, nm
To examine the reaction barriers, we studied the transition states
for the concerted mechanism of the [2π + 2π] cycloaddition of
thyminyl derivatives. It was shown recently that the photo-
induced cycloaddition of thymine followed the concerted mech-
anism rather than the biradical pathway.31 The latter was found
to dominate during the thermal cycloaddition.31 The photo-exci-
tation of thymine results in a barrierless, non-radiative decay of a
singlet excited state with a low-lying conical intersection
between the ground state (S0) and excited state (S1) energetic
profiles. At the S0/S1 conical intersection the system displays a
fair degree of puckering around the CvC bond leading to a for-
mation of the ground-state cyclobutane dimer. It was established
that the concerted mechanism of cycloaddition could be accu-
rately treated with a single-reference method, such as second-
order perturbation theory, MP2.62 For this reason the transition
states for the concerted mechanism in the current study were
located at the B3LYP/6-31G(d) level theory and the reaction bar-
riers were then improved with MP2/cc-pVTZ. All of the tran-
sition states resembled the thymine dimer at the conical
intersection in ref. 31 and displayed one imaginary frequency
corresponding to the formation of the cyclobutane ring. Four
possible configurations of the transition states, such as TA, TS,
CA and CS, were calculated for four monomers: methyl acetate
(3), propanoic acid (6), methyl propanoate (5) and propanamide
(7) (see Fig. 6). The MP2 improved reaction barriers are pres-
ented in Table 2.
Acetate (3)
TA
CA
TS
CS
TA
CA
TS
CS
TA
CA
TS
CS
TA
CA
TS
CS
207.7
204.9
258.6
285.0
211.9
214.4
282.0
318.7
216.5
225.7
300.9
323.5
170.7
205.1
273.3
290.5
574.2
582.0
461.2
418.5
562.8
556.3
422.9
374.2
550.9
528.3
396.4
368.7
698.6
581.6
436.4
410.6
Propanoate (5)
Propanoic acid (6)
Propanamide (7)
Therefore, regardless of the transition state configuration, the
ring closure to form cyclobutane dimer should occur for all four
monomers. This conclusion was confirmed by the solution-phase
photo-dimerisation of 7, in which all four photo-dimers were
obtained in low yield (<7.5% total yield) and with reduced
stereospecificity (see ESI†). From the solution-phase irradiation,
the TS dimer of 7 was obtained with 8% specificity, whereas it
was obtained with 80% specificity from the solid-state
irradiation.
Apart from the CA configuration the transition states depict
the asynchronous mechanism of the ring closure, with one
forming C–C bond being longer than the other. For more details
see the ESI.† Structurally, it requires a significant change in the
ring and a perfect alignment of the CvC bonds with about 2 Å
separation for the ring closure to occur. The monomers should
not only align with the CvC bonds but they should also move
closer to allow for cycloaddition. As a result, in the crystal struc-
ture we rely on the specific arrangement of the thyminyl rings,
such that when the thyminyl derivates become excited upon UV
exposure, they can move into the position that is most preferable
for the ring closure to occur. This puts a constraint on the
packing arrangement and the crystal structure should allow for a
fair amount of flexibility for the thyminyl rings to adopt the pre-
ferred position. Out of the six monomers studied, three were
crystallised in the CS slipped and displaced configuration that
would not be ideal for the ring closure reaction and would
require a high degree of rearrangement for the thyminyl rings. It
can be noted that if the thyminyl rings are involved in other
intermolecular interactions rather than π–π stacking, the flexi-
bility of the rings in the crystalline state is rather limited and
photo-dimerisation becomes highly unlikely. Thus, it appears
that the packing arrangement plays a very important role in deter-
mining the extent of photo-dimerisation. The only successfully
dimerised monomer crystal, propanamide 7, displayed the TS
criss-cross configuration (Fig. 1, Fig. 3d). In this arrangement
the thyminyl rings would only need to slightly reorient for
[2π + 2π]-cycloaddition to proceed. To summarise, as reactant
topochemistry is the determinant factor in the outcome of any
[2π + 2π] cycloaddition in the crystalline state, it appears that
intermolecular interactions between monomers play a significant
Reaction barriers for the same configuration vary only slightly
by varying the N1 substituent. The TA and CA configurations
produce the lowest reaction barriers. Energetically, the barriers
vary from as low as 170 kJ mol−1 for the TA configuration in
propanamide 7 to 323 kJ mol−1 for the CS configuration in pro-
panoic acid 6. Although the energetic difference in these barriers
is quite significant, re-calculation to wavelengths shows that the
barriers are well within the energy of 302 nm UV light.
Fig. 6 Optimised structures of the transitions states for the concerted
mechanism of photo-induced [2 + 2] cycloaddition of propanamide in
four possible configurations: TA, CA, TS and CS. All distances are
given in Å.
This journal is © The Royal Society of Chemistry and Owner Societies 2012
Photochem. Photobiol. Sci., 2012, 11, 1938–1951 | 1945

View Article Online
role in accommodating the favourable π–π stacking to allow for
photo-dimerisation. Therefore, the effect of the substituents on
the packing arrangements of the N1 derivatives needs to be
thoroughly investigated.
initial geometries and three of the monomers, 2, 6 and 7,
optimised to other intermolecular interactions shown
under special cases 1 (SC1) in Table 3. Special cases 2 (SC2)
shown in the last column represent intermolecular
interactions observed in the crystal structures that do not fit the
description of the four traditional types of interactions as
depicted in Fig. 3.
Substituent effects on strength of intermolecular interactions
observed in the crystal structures
Analysis of Table 3 reveals an interesting fact that for each
monomer, regardless of the nature of the N1 derivative, the π–π
stacking in the preferable TA arrangement and the classic WC
hydrogen bonding differ only by a couple of kJ mol−1, which is
well within the systematic error of the calculations. The only
exception is the dimer of acetic acid 2, where the Watson and
Crick hydrogen bonding is preferred by about 11 kJ mol−1 and
might therefore prevent the thyminyl rings from stacking. There-
fore, it is not surprising to find thyminyl rings in the crystal
structure of acetic acid 2 in the CS slipped and displaced confor-
mation that does not favour cycloaddition. Out of the six mono-
mers, propanamide 7 produces the strongest π–π stacking and
Secondly, to examine the strength of the various intermolecular
interactions observed in the crystal structures of 2–7, the specific
intermolecular interactions were extracted from the crystal data
and re-optimised at the B3LYP level of theory. The improved
MP2 binding energies of the four types of intermolecular inter-
actions (as shown in Fig. 3) as well as special cases (as shown in
Fig. 7) are presented in Table 3. The first column of Table 3
shows the strength of the π–π stacking interactions modelled
with respect to the preferable TA arrangement of thyminyl rings
found in the methyl propanoate crystal structure. The geometry
optimisations confirmed this arrangement for all but acetamide
that preferred to optimise to the WC style hydrogen bonding
instead (see Fig. 7). Similarly, the original π–π stacking arrange-
ment found in the crystal structures apart from 3 were taken as
WC hydrogen bonding interactions in excess of >20 kJ mol−1
.
In the case of acetamide 4, the strength of the π–π stacking and
WC HB interactions is similar to those of other monomers due
to intramolecular hydrogen bonding between the carbonyl
oxygen of the ring and the hydrogen atom of the amide group,
preventing the formation of stronger inter-ring interactions. Due
to the flexibility of the ethyl linker in propanamide 7 this intra-
molecular hydrogen bonding is absent in the monomer itself,
thus resulting in a stronger WC hydrogen bond between the
rings. It has to be noted that the thyminyl rings in the π–π stack-
ing dimers (presented in the first and third columns of Table 3)
are separated by about 3.7 Å on average regardless of the N1
derivative and the type of stacking. The calculations performed
in gas phase usually produce shorter intermolecular distances
(by about 0.5 Å)63 than those observed in the crystal structure
due to exclusion of interactions with the bulk of the crystal. This
type of interaction decays very fast with distance (∼R−6) and at
the separation of >5 Å could even become negligible.64 For
monomers with functionalities allowing for additional hydrogen
bonding, such as carboxylic acids (2 and 6) and amides (4 and
7), inter-chain and chain–ring interactions are also relatively
strong. The chain–ring interaction ranges from −48 kJ mol−1 for
propanoic acid 6 to as much as −83 kJ mol−1 for propanamide
7, making it the strongest interaction among the dimers studied.
The inter-chain interaction in the dimer of acetic acid 2 is not
only the strongest inter-chain interaction among the six mono-
mers but also the strongest intermolecular interaction for acetic
acid 2 overall.
The results on the MP2 binding energies indicate that for each
individual monomer there is no significant variation in strength
of the four main types of intermolecular interactions observed in
the crystal structures. Therefore, all four types are in competition
with each other and could potentially be present in approxi-
mately equal amounts in the crystals of the six monomers
studied. In this study “ideal” interactions between two monomer
units were considered. One should keep in mind that each
monomer is usually involved in multiple intermolecular inter-
actions at once, thus further strengthening some intermolecular
interactions and weakening the others. The molecular arrange-
ment in the crystal is thus a subtle interplay between not only
Fig. 7 Optimised structures of the special cases found in the crystal
structures. All distances are given in Å.
1946 | Photochem. Photobiol. Sci., 2012, 11, 1938–1951
This journal is © The Royal Society of Chemistry and Owner Societies 2012

View Article Online
Table 3 Summary of the binding energies for intermolecular interactions observed in 2–7
π–π stacking TAa
WC HB
π–π stacking crystal
Inter-chain HB
Chain–ring HB
SC1b
SC2c
2
3
4
−43.5
−54.0
SC1
−54.1
−53.8
−59.7
−53.8d
−55.4
−50.7
−75.6d
SC1
N/A
−50.6
−69.4
N/A
−53.7
−58.5
−47.0
−50.3
−47.5
−76.1
−41.1
5
6
7
−53.9
−49.6
−73.0
−53.9
SC1
SC1
N/A
N/A
−55.6
N/A
−48.3
−83.1
−36.4
−71.3
−38.5
a Structures adapted to the TA configuration of the propanoate dimer observed in the crystal structure. b Optimized structures, for which the
arrangement in the crystal produced a different configuration after geometry optimisation. These structures are denoted here as special cases 1 (SC1).
c Special cases of the observed inter-molecular interactions between two thyminyl monomers that do not fit the description of the other four groups.
These structures are denoted here as special cases 2 (SC2). d Anti-Watson and Crick HB as observed in the crystal structure.
the strength of individual intermolecular interactions, but also
cooperativity effects of certain interactions, e.g. hydrogen
bonding,65 that in turn, might stabilise the crystal lattice to a
much greater extent.
A thorough analysis of the crystal structures was performed,
which concentrated on the number of specific intermolecular
interactions per the same number of monomer molecules (ESI†).
Acetic acid 2, acetate 3 and propanoic acid 6 displayed an equal
Fig. 8 N3–N3 butyl-linked derivatives.
amount of WC and chain–ring hydrogen bonds per 8 monomer
molecules, which was double the amount of the π–π stacking
interactions. These observations are in agreement with the results
obtained concerning the strength of these intermolecular inter-
actions. In acetamide 4 anti-WC bonds, chain–chain and chain–
ring interactions equally contributed to the packing arrangement
Alkylation at the N3 position to eliminate Watson and Crick
hydrogen bonding, and its influence on the crystal packing and
photo-reactivity of the bis-thyminyl compounds
To address the strategies proposed in the computational section
our research goal was to synthesise symmetrical bis-thyminyl
derivatives (8–10, Fig. 8) by bridging two of the thyminyl units
at the N3 position using an n-butyl spacer. This served two pur-
poses. Firstly, N3–N3-alkyl bridging would effectively block the
Watson and Crick hydrogen bonding, and; secondly, the
additional thyminyl unit could potentially make the monomer
system more rigid and hence, more susceptible to the preferred
π–π stacking interactions.
The synthesis, crystallisation, and structure solutions of mono-
mers 8 and 10 were described previously.33 The bis-propanoic
acid monomer, 9, was obtained by base hydrolysis of the bis-
methyl propanoate monomer, 8. Single crystals of bis-propanoic
9, suitable for photo-reactivity and X-ray analysis experiments
were obtained by the rapid acidification of the hydrolysis reac-
tion mixture.
of the monomer, whereas the π–π stacking constituted only 14%
of the overall interactions. In these four cases the separation
between the rings was too long, >4.9 Å, making the π–π stacking
interaction too weak to drive the [2π + 2π] cycloaddition. In
methyl propanoate 5, the Watson and Crick HB dominated the
crystal structure due to unavailability of the ester group to
partake in other specific interactions. Although the π–π stacking
in the preferable TA arrangement was still observed, the separ-
ation of 4.4 Å between the rings was quite long, potentially ren-
dering this interaction relatively weak. In accordance with the
computational results, a network of the chain–ring and inter-
chain interactions was observed in propanamide 7. The coopera-
tivity of these specific interactions also allowed for a network of
the TS criss-cross π–π packing to occur at a favourable separ-
ation of about 3.7 Å, thus making the latter a very strong
interaction.
To this end, we can conclude that in order to encourage the
formation of strong π–π stacking interactions between thyminyl
rings in the preferable TA arrangement the energetic competition
with specific chain–ring and inter-chain interactions should be
eliminated. Three proposed strategies to achieve this goal
were: blocking the N3 nitrogen of the thyminyl ring from
forming the Watson and Crick style hydrogen bonding;
including functional groups on the N1 derivatives that are
less susceptible to hydrogen bonding (i.e. such as methyl
propanoate); and finally, excluding the intramolecular inter-
actions between the N1 chains and the ring by incorporating
longer alkyl linkers, which will permit greater flexibility of the
chains.
Photo-irradiation study of bis-thyminyl derivatives, 8–10
To examine the photo-reactivity of the bis-thyminyl monomer
crystals (8–10), samples of each compound were irradiated
(0.30 kJ cm−2), and the degrees of conversion to cyclobutane
were determined by subsequent 1H NMR analysis of the ir-
1
radiated samples. The calculated H NMR yields summarised in
Table 4, were determined from the ratio of C5–CH3 methyl
protons of non-reacted thyminyl groups to the cyclobutane C5–
CH3 methyl protons. Referring to Table 4, it was evident that the
bis-propanoate 8 and bis-propanoic 9 monomers underwent
This journal is © The Royal Society of Chemistry and Owner Societies 2012
Photochem. Photobiol. Sci., 2012, 11, 1938–1951 | 1947

View Article Online
Table 4 Interactions observed in the bis-thyminyl crystal structures when N3 is alkylated. All distances are given in Å
System, space
group
1H NMR
yield
Olefinic separation
distance (Å)
Type of π–π
stacking
WC
HB
Cmpd.
Inter-chain HB
None
Chain–ring HB
8
Monoclinic,
96
4.21 Å
TA, sandwich
NA
None
P21/c 33
ˉ
9
10
Triclinic, P1
81
—
4.19 Å
4.77 Å
TA, sandwich
CS, slipped and
displaced
NA
NA
None
None
CONHH⋯OvC2,
2.01 Å
Tetragonal, P42/n 33
CONHH⋯OvCNH2,
1.96 Å
Fig. 9 Hydrogen bonding and π–π stacking interactions in the crystal structures of the N3-blocked bis-thyminyl monomers (8–10).
extensive photo-conversion to the corresponding cyclobutane
compounds (96% and 81%, respectively). Conversely, the bis-
propanamide monomer (10) was photo-stable. To understand the
structural reasons contributing to the successful photo-conver-
sion of monomer crystals of 8 and 9, but not 10, the crystal
structures of each monomer were examined.
As mentioned, the crystal structures of monomers 8 and 10
were discussed in a recent report.33 However, the crystal struc-
ture of the new monomer, bis-propanoic acid 9, will be briefly
discussed here. For convenience, some of the key interactions
observed in the crystal structures of monomers 8–10 are pre-
sented in Table 4.
Due to the weakly diffracting nature of crystals obtained for 9,
the structure was determined from a single-crystal diffraction
experiment performed on the micro-crystallography beamline at
the Australian Synchrotron.
It was expected that inter-chain hydrogen-bond driven assem-
bly of monomers would occur in the butyl-linked bis(propanoic
acid) (9) crystals due to the free carboxylic moieties. However,
instead of the anticipated intermolecular hydrogen bonding
between monomer molecules, the interactions instead occurred
between solvent and monomer molecules. Water molecules
present in the crystal lattice from crystallisation actually provided
sufficient hydrogen bonding sites to entirely eliminate inter-
monomer hydrogen-bonding, and subsequently permit similar
trans–anti thyminyl ring-stacking, to that observed in the bis-
propanoate monomer structure, 8.
The overall structure of 9 was stabilised by three unique
hydrogen bonds between the monomer molecules and lattice
included water molecules (Fig. 9). Each of the propanoic acid
groups bonded with the O of a water molecule (COOH⋯O(W),
1.70 Å), while the hydrogen atoms of the water molecule partici-
pate in hydrogen-bonds with the thyminyl carbonyl oxygen
atoms (H(W)⋯OvC4, 1.81 Å; and H(W)⋯OvC2, 1.93 Å).
The closest thyminyl pairing arose from TA stacked rings, whose
olefinic groups were separated by 4.19 Å.
Referring to Table 4, expectedly the alkyl bridging between
the N3 positions eliminated Watson and Crick hydrogen bonding
interactions entirely. Of further interest is that this modification
also eliminated inter-chain and chain–ring hydrogen bonding
between the bis-propanoic acid molecules (9), due to hydrogen
bonding between monomer molecules and lattice included water
molecules. The photo-reactive monomer crystals of 8 and 9, also
exhibited similar TA sandwich stacking (Fig. 9) between the
proximity-related thyminyl rings; and in both cases, this was
accompanied by the desired short olefinic separation distances of
1948 | Photochem. Photobiol. Sci., 2012, 11, 1938–1951
This journal is © The Royal Society of Chemistry and Owner Societies 2012

View Article Online
4.21 and 4.19 Å, respectively. Not surprisingly, in the bis-propa-
namide monomer crystals (10), inter-chain hydrogen-bonding
(NH⋯OvC, 1.96 Å) and chain–ring hydrogen bonding
(amide NH⋯OvC2, 2.01 Å) gave rise to proximity-related
CS thyminyl stacks that were slipped and displaced (Fig. 9). The
displacement observed between the thyminyl moieties of
molecules of 10 resulted in large olefinic separation distances of
4.77 Å that were inappropriate for [2π + 2π]-cycloaddition
reactions.
It was found from this analysis that photo-reactive bis-thymi-
nyl monomer crystals were obtained when Watson and Crick,
inter-chain and chain–ring interactions were eliminated (i.e. in
the cases of monomer crystals 8 and 9). In both cases, the
absence of these intermolecular interactions facilitated the prefer-
able TA arrangement of thyminyl rings and close olefinic separ-
ation distances of around 4.2 Å.
two thyminyl rings. The rigidity of the rings led to favourable
TA π–π stacking in 8 and 9, whereas in 10 the inter-chain inter-
actions between amide groups dominated the crystal structure.
The best reversible photo-polymerisation was obtained in 8,
whose crystal structure was driven by the preferable TA arrange-
ment of thyminyl rings in the absence of other specific inter-
actions. In the case of 9, the carboxylic groups were blocked
from forming inter-chain interactions due to the presence of
water molecules acting as stabilising agents, thereby leading to
some degree of oligomerisation. In order to obtain more numer-
ous photo-reactive bis-thyminyl derivatives, our future investi-
gations will focus on varying the N3-bridge within a constant
bis-methyl propanoate architecture. In doing so, it should also be
possible to gain insight into the influence of different bridging
species on the crystal packing and photo-reactivity of the result-
ing compounds. More crystallographic and quantum chemical
studies will also be required to assemble a library of thyminyl
derivatives that are amenable to chemical bridging and that also
exhibit favourable π–π stacking arrangements for [2π + 2π]-
cycloaddition reactions.
Conclusions
For the N1 derivatives studied (2–7), the calculated reaction bar-
riers for the concerted mechanism of photo-induced [2π + 2π]
cycloaddition indicated energetic preference for the trans–anti
and cis–anti arrangements of thyminyl rings. Although the reac-
tion barriers were quite high (>170 kJ mol−1), irradiation using
302 nm UV light provided sufficient amount of energy for their
photo-dimerisation to occur for all four types of arrangements
(TA, CA, TS and CS). The solution-phase irradiation of 7 sup-
ported this notion since the photo-reactions yielded mixtures of
the photo-dimers in relatively low yields <10%, and with
reduced stereospecificity.
Notes and references
1 A. Jayaraman, V. Balasubramaniam and S. Valiyaveettil, Self-assembly of
pentaphenol adducts: Formation of 3D network and ladder-type supra-
molecular structures in the solid state, Cryst. Growth Des., 2006, 6, 636–
642.
2 T. Friščić and L. R. MacGillivray, Cyclophanes and ladderanes: molecular
targets for supramolecular chemists, Supramol. Chem., 2005, 17, 47–51.
3 X. Gao, T. Friščić and L. R. MacGillivray, Supramolecular construction
of molecular ladders in the solid state, Angew. Chem., Int. Ed., 2004, 43,
232–236.
The transition state analyses indicate that for the [2π + 2π]-
cycloaddition to occur an appropriate alignment of the CvC
bonds should be achieved. Out of the six monomers studied
irradiation of the crystalline sample of the propanamide 7
resulted in dimerisation and produced mostly the trans–syn
cyclobutane dimer in ca. 44% yield. The crystal structure of this
particular derivative showed suitable π–π stacking between the
thyminyl rings. Although propanoate displayed the preferred TA
arrangement for the π–π stacking, the separation between the
rings was relatively long for the cycloaddition to take place.
The strengths of the four specific interactions, such as π–π
stacking, Watson and Crick, inter-chain and chain–ring hydrogen
bonding, were subjected to rigorous computational studies with
the view of quantifying the strength of each individual inter-
action. Strikingly, only a slight variation in the strength of these
interactions was found for each monomer, thus highlighting the
fact that the packing arrangement in the crystal was due to the
interplay between the strength of the interaction and cooperativ-
ity effects. As a result, three strategies were proposed to encou-
rage the formation of preferable π–π stacking arrangement: (1)
Blocking the N3 nitrogen of the thyminyl ring, and thereby elim-
inating WC HB; (2) Inclusion of functional groups on the N1
derivatives that are less susceptible to hydrogen bonding; and (3)
Exclusion of intramolecular interactions between the N1 chains
and the ring by incorporating longer alkyl linkers.
4 G. W. Coates, A. R. Dunn, L. M. Henling, J. W. Ziller, E. B. Lobkovsky
and R. H. Grubbs, Phenyl-perfluorophenyl stacking interactions: topo-
chemical [2 + 2] photodimerization and photopolymerization of olefinic
compounds, J. Am. Chem. Soc., 1998, 120, 3641–3649.
5 M. Hasegawa, Four-center type photopolymerization in the solid state,
Int. Chem. Eng., 1971, 11, 177–188.
6 P. Klán and J. Wirz, in Photochemistry of Organic Compounds, John
Wiley and Sons Ltd, 2009, pp. 227–453.
7 A. Sergeiko, V. V. Poroikov, L. O. Hanu and V. M. Dembitsky, Cyclobu-
tane-containing alkaloids: origin, synthesis, and biological activities,
Open Med. Chem. J., 2008, 2, 26–37.
8 G. M. J. Schmidt, Photodimerizations in the solid state, Pure Appl.
Chem., 1971, 27, 647–678.
9 G. J. Fisher and H. E. Johns, Ultraviolet photochemistry of thymine in
aqueous solution, Photochem. Photobiol., 1970, 11, 429–444.
10 X. Mei, S. Liu and C. Wolf, Template-controlled face-to-face stacking of
olefinic and aromatic carboxylic acids in the solid state, Org. Lett., 2007,
9, 2729–2732.
11 T. Caronna, R. Liantonio, T. A. Logothetis, P. Metrangolo, T. Pilati and
G. Resnati, Halogen bonding and π⋯π stacking control reactivity in the
solid state, J. Am. Chem. Soc., 2004, 126, 4500–4501.
12 Q. Chu, D. C. Swenson and L. R. MacGillivray, A single-crystal-to-
single-crystal transformation mediated by argentophilic forces converts a
finite metal complex into an infinite coordination network, Angew.
Chem., Int. Ed., 2005, 44, 3569–3572.
13 L. T. Ni, M. Nagarathinam and J. J. Vittal, Topochemical photodimeriza-
tion in the coordination polymer [{(CF3CO2)(μ-O2CCH3)Zn}2(μ-bpe)2]n
through single-crystal to single-crystal transformation, Angew. Chem., Int.
Ed., 2005, 44, 2237–2241.
14 G. S. Papaefstathiou, C. Milios and L. R. MacGillivray, A 2D metal–
organic framework with two different rhombus-shaped cavities: a rare
example of a (4,4)-net with alternating metal and organic nodes, Micro-
porous Mesoporous Mater., 2004, 71, 11–15.
15 M. Yoshizawa, Y. Takeyama, T. Kusukawa and M. Fujita, Cavity-directed,
highly stereoselective [2 + 2] photodimerization of olefins within self-
Elimination of unfavourable Watson and Crick hydrogen
bonding in the crystal was successfully achieved in bis-thyminyl
derivatives 8–10, upon installation of an n-butyl-bridge between
This journal is © The Royal Society of Chemistry and Owner Societies 2012
Photochem. Photobiol. Sci., 2012, 11, 1938–1951 | 1949

View Article Online
assembled coordination cages, Angew. Chem., Int. Ed., 2002, 41, 1347–
1349.
40 L. J. Barbour, Xseed, University of Missouri, Columbia, 1999.
41 L. Kosynkina, W. Wang and T. C. Liang, A convenient synthesis of chiral
peptide nucleic acid (PNA) monomers, TetrahedronLett., 1994, 35,
5173–5176.
42 M. C. Liu, W. Feng and J. C. Ding, 1-(Carboxymethyl)thymine,
Acta Crystallogr., Sect. E: Struct. Rep. Online, 2004, 60, o1611–
o1612.
16 L. R. MacGillivray, Organic synthesis in the solid state via hydrogen-
bond-driven self-assembly, J. Org. Chem., 2008, 73, 3311–3317.
17 L. R. MacGillivray, J. L. Reid and J. A. Ripmeester, Supramolecular
control of reactivity in the solid state using linear molecular templates
[4], J. Am. Chem. Soc., 2000, 122, 7817–7818.
18 L. R. MacGillivray, J. L. Reid, J. A. Ripmeester and G. S. Papaefstathiou,
Toward a reactant library in template-directed solid-state organic syn-
thesis: reactivity involving a monofunctional reactant based on a stilba-
zole, Ind. Eng. Chem. Res., 2002, 41, 4494–4497.
43 F. Debaene, J. A. Da Silva, Z. Pianowski, F. J. Duran and N. Winssinger,
Expanding the scope of PNA-encoded libraries: divergent synthesis of
libraries targeting cysteine, serine and metallo-proteases as well as tyro-
sine phosphatases, Tetrahedron, 2007, 63, 6577–6586.
19 I. G. Ovchinnikov, O. V. Fedorova, P. A. Slepukhin and G. L. Rusinov,
Photoinduced template synthesis of cyclobutane-containing crown ether,
Russ. Chem. Bull., 2008, 57, 212–214.
44 F. H. Allen, C. A. Baalham, J. P. M. Lommerse and P. R. Raithby, Carbo-
nyl–carbonyl interactions can be competitive with hydrogen bonds, Acta
Crystallogr., Sect. B: Struct. Sci., 1998, 54, 320–329.
20 W. Herrmann, M. Schneider and G. Wenz, Photochemical synthesis of
polyrotaxanes from stilbene polymers and cyclodextrins, Angew. Chem.,
Int. Ed. Engl., 1997, 36, 2511–2514.
21 W. Herrmann, S. Wehrle and G. Wenz, Supramolecular control of the
photochemistry of stilbenes by cyclodextrins, Chem. Commun., 1997,
1709–1710.
45 J. Spychała, A convenient synthesis of monocationic 1-[(cyclic amidino)-
methyl]thymines, Synth. Commun., 1997, 27, 3431–3440.
46 T. Borowiak, G. Dutkiewicz and J. Spychała, Molecular packing
in (5-methyl-2,4-dioxo-1,2,3,4-tetrahydropyrimidin-1-yl) acetamide,
Acta Crystallogr., Sect. C: Cryst. Struct. Commun., 2006, 62, o344–
o345.
22 M. Pattabiraman, A. Natarajan, L. S. Kaanumalle and V. Ramamurthy,
Templating photodimerization of trans-cinnamic acids with cucurbit[8]-
uril and γ-cyclodextrin, Org. Lett., 2005, 7, 529–532.
23 K. S. S. P. Rao, S. M. Hubig, J. N. Moorthy and J. K. Kochi, Stereo-
selective photodimerization of (E)-stilbenes in crystalline γ-cyclodextrin
inclusion complexes, J. Org. Chem., 1999, 64, 8098–8104.
24 Y. B. Ito Borecka, G. Olovsson, J. Trotter and J. R. Scheffer, Control of
the solid-state photodimerization of some derivatives and analogs of
trans-cinnamic acid by ethylenediamine, Tetrahedron Lett., 1995, 36,
6087–6090.
47 N. Robert, C. Hoarau, S. Célanire, P. Ribéreau, A. Godard, G. Quéguiner
and F. Marsais, Efficient and fast Heck vinylation of 2-bromo-6-methyl
pyridines with methylacrylate. Application to the synthesis of 6-methyl
cyclopenta[b]pyridinone, Tetrahedron, 2005, 61, 4569–4576.
48 S. Boncel and K. Walczak, Novel acyclic amide-conjugated nucleosides
and their analogues, Nucleosides, Nucleotides Nucleic Acids, 2009, 28,
103–117.
49 K. Kizawa, A. Takenaka and Y. Sasada, Structure of 1-(2-carbamoyl-
ethyl)thymine, Acta Crystallogr., Sect. C: Cryst. Struct. Commun., 1988,
44, 535–537.
25 J. N. Moorthy and K. Venkatesan, Unusual photobehaviour of
crystalline 7-methoxy-4-methylcoumarin, J. Mater. Chem., 1992, 2, 675–
676.
26 M. Pattabiraman, A. Natarajan, R. Kaliappan, J. T. Mague and
V. Ramamurthy, Template directed photodimerization of trans-1,2-bis-
(n-pyridyl)-ethylenes and stilbazoles in water, Chem. Commun., 2005,
4542–4544.
27 Y. Inaki and H. Hiratsuka, Reversible photo-cross-linking of polyvinyl
alcohol containing thymine, J. Photopolym. Sci. Technol., 2000, 13, 739–
744.
28 T. Carell and R. Epple, Repair of UV light induced DNA lesions: a com-
parative study with model compounds, Eur. J. Org. Chem., 1998, 1245–
1258.
29 G. M. Blackburn and R. J. H. Davies, The structure of thymine photo-
dimer, J. Chem. Soc. C, 1966, 2239–2244.
30 W. J. Schreier, T. E. Schrader, F. O. Koller, P. Gilch, C. E. Crespo-Her-
nández, V. N. Swaminathan, T. Carell, W. Zinth and B. Kohler, Thymine
dimerization in DNA is an ultrafast photoreaction, Science, 2007, 315,
625–629.
31 M. Boggio-Pasqua, G. Groenhof, L. V. Schäfer, H. Grubmüller and
M. A. Robb, Ultrafast deactivation channel for thymine dimerization,
J. Am. Chem. Soc., 2007, 129, 10996–10997.
32 L. Blancafort and A. Migani, Modeling thymine photodimerizations in
DNA: mechanism and correlation diagrams, J. Am. Chem. Soc., 2007,
129, 14540–14541.
50 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb,
J. R. Cheeseman, J. A. Montgomery Jr., T. Vreven, K. N. Kudin,
J. C. Burant, J. M. Millam, S. S. Iyengar, J. Tomasi, V. Barone,
B. Mennucci, M. Cossi, G. Scalmani, N. Rega, G. A. Petersson,
H. Nakatsuji, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa,
M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X. Li,
J. E. Knox, H. P. Hratchian, J. B. Cross, V. Bakken, C. Adamo,
J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin,
R. Cammi, C. Pomelli, J. W. Ochterski, P. Y. Ayala, K. Morokuma,
G. A. Voth, P. Salvador, J. J. Dannenberg, V. G. Zakrzewski, S. Dapprich,
A. D. Daniels, M. C. Strain, O. Farkas, D. K. Malick, A. D. Rabuck,
K. Raghavachari, J. B. Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul,
S. Clifford, J. Cioslowski, B. B. Stefanov, A. L. G. Liu, P. Piskorz,
I. Komaromi, R. L. Martin, D. J. K. Fox, T. M. A. Al-Laham, C. Y. Peng,
A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W. Chen,
M. W. Wong, C. Gonzalez and J. A. Pople, Gaussian, Gaussian Inc.,
Wallingford CT, 2009.
51 A. D. Becke, Density-functional exchange-energy approximation with
correct asymptotic behavior, Phys. Rev. A: At., Mol., Opt. Phys., 1988,
38, 3098–3100.
52 C. Lee, W. Yang and R. G. Parr, Development of the Colle–Salvetti corre-
lation-energy formula into a functional of the electron density, Phys. Rev.
B, 1988, 37, 785–789.
53 A. D. Becke, Density-functional thermochemistry. III. The role of exact
exchange, J. Chem. Phys., 1993, 98, 5648–5652.
33 P. Johnston, C. Braybrook and K. Saito, Topochemical photo-
reversible polymerization of a bioinspired monomer and its recovery and
repolymerization after photo-depolymerization, Chem. Sci., 2012, 3,
2301–2306.
34 D. G. Truhlar, B. C. Garrett and S. J. Klippenstein, Current status of tran-
sition-state theory, J. Phys. Chem., 1996, 100, 12771–12800.
35 T. M. McPhillips, S. E. McPhillips, H. J. Chiu, A. E. Cohen,
A. M. Deacon, P. J. Ellis, E. Garman, A. Gonzalez, N. K. Sauter,
R. P. Phizackerley, S. M. Soltis and P. Kuhn, Blu-Ice and the distributed
control system: software for data acquisition and instrument control at
macromolecular crystallography beamlines, J. Synchrotron Radiat., 2002,
9, 401–406.
54 P. C. Hariharan and J. A. Pople, The influence of polarization functions
on molecular orbital hydrogenation energies, Theor. Chim. Acta, 1973,
28, 213–222.
55 C. Møller and M. S. Plesset, Note on an approximation treatment for
many-electron systems, Phys. Rev., 1934, 46, 618–622.
56 J. Thom and H. Dunning, Gaussian basis sets for use in correlated mo-
lecular calculations. I. The atoms boron through neon and hydrogen,
J. Chem. Phys., 1989, 90, 1007–1023.
57 Y. Inaki, E. Mochizuki, N. Yasui, M. Miyata and Y. Kai, Crystal structure
and photodimerization of 1-alkylthymine: effect of long alkyl chain
on isomer ratio of photodimer, J. Photopolym. Sci. Technol., 2000, 13,
177–182.
36 W. J. Kabsch, J. Appl. Crystallogr., 1993, 26, 795.
37 R. Blessing, Outlier treatment in data merging, J. Appl. Crystallogr.,
1997, 30, 421–426.
38 G. Sheldrick, Phase annealing in SHELX-90: direct methods for larger
structures, Acta Crystallogr., Sect. A: Found. Crystallogr., 1990, 46, 467–
473.
39 G. M. Sheldrick, SHELXL-97, University of Göttingen, Göttingen,
1997.
58 E. Mochizuki, N. Yasui, Y. Kai, Y. Inaki, W. Yuhua, T. Saito, N. Tohnai
and M. Miyata, Crystal structure of long alkyl 3-(thymin-1-yl)propio-
nates: style of hydrogen bonding and dependence on the alkyl chain
length, Bull. Chem. Soc. Jpn., 2001, 74, 193–200.
59 M. Monajjemi and B. Chahkandi, Theoretical investigation of hydrogen
bonding in Watson–Crick, Hoogestein and their reversed and other
models: comparison and analysis for configurations of adenine–thymine
base pairs in 9 models, THEOCHEM, 2005, 714, 43–60.
1950 | Photochem. Photobiol. Sci., 2012, 11, 1938–1951
This journal is © The Royal Society of Chemistry and Owner Societies 2012

View Article Online
60 N. Tohnai, Y. Inaki, M. Miyata, N. Yasui, E. Mochizuki and Y. Kai,
Photoreactive and inactive crystals of 1-alkylthymine derivatives, Bull.
Chem. Soc. Jpn., 1999, 72, 851–858.
61 J. Šponer, J. Leszczynski and P. Hobza, Hydrogen bonding, stacking and
cation binding of DNA bases, THEOCHEM, 2001, 573, 43–53.
62 B. Szefczyk, T. Andruniow and W. A. Sokalski, Ab initio multireference
study of Hetero-Diels–Alder reaction of buta-1,3-diene with alkyl
glyoxylates, J. Mol. Model., 2008, 14, 727–733.
63 G. B. Deacon, C. M. Forsyth, P. C. Junk, E. I. Izgorodina, T. J. Ness,
G. Meyer and I. Pantenburg, A supramolecular twist to the structures
of bis(polyfluorophenyl)mercurials, CrystEngComm, 2011, 13, 88–
92.
64 A. J. Stone, The Theory of Intermolecular Forces, Clarendon Press,
Oxford, 1997.
65 D. Hadži, Theoretical treatment of hydrogen bonding, John Wiley and
Sons, 1997.
This journal is © The Royal Society of Chemistry and Owner Societies 2012
Photochem. Photobiol. Sci., 2012, 11, 1938–1951 | 1951