The molecular design, synthesis and photochromic properties of spirooxazines containing a permanent azo (hydrazone) chromophore

Article abstract of DOI:10.1016/j.dyepig.2013.02.017

A range of molecular modelling techniques has been applied to predict optimized geometrical conformations and energies of the ring-closed oxazine form and ring-opened merocyanine forms of three azospirooxazines. A comparison of the relative stabilities of the possible isomers of the merocyanines was carried out by applying molecular mechanics (augmented MM2), while the prediction of the potential for photochromic behaviour was made by comparison of the heats of formation calculated by AM1. Synthetic routes were explored and as a result two azospirooxazines were isolated, one of which showed interesting and unusual photochromic behaviour. The instability of this compound towards purification by chromatography was investigated. The decomposition product was isolated and characterized with the aid of single crystal x-ray crystallography. The correlation between predicted and observed photochromic behaviour is discussed.

Full text of DOI:10.1016/j.dyepig.2013.02.017

Dyes and Pigments 98 (2013) 263e272
Contents lists available at SciVerse ScienceDirect
Dyes and Pigments
journal homepage: www.elsevier.com/locate/dyepig
The molecular design, synthesis and photochromic properties
of spirooxazines containing a permanent azo (hydrazone)
chromophore
,
c
,
a
a
a
Robert M. Christie a , Keith M. Morgan , Ayesha Rasheed , Mohanad Aldib ,
*
Georgina Rosair b
a School of Textiles and Design, Heriot-Watt University, Scottish Borders Campus, Galashiels TD1 3HF, Scotland, UK
b School of Engineering and Physical Sciences, Heriot-Watt University, Edinburgh EH14 4AS, Scotland, UK
c Department of Chemistry, Faculty of Science, King Abdulaziz University, Jeddah 21589, Saudi Arabia
a r t i c l e i n f o
a b s t r a c t
Article history:
A range of molecular modelling techniques has been applied to predict optimized geometrical confor-
mations and energies of the ring-closed oxazine form and ring-opened merocyanine forms of three
azospirooxazines. A comparison of the relative stabilities of the possible isomers of the merocyanines
was carried out by applying molecular mechanics (augmented MM2), while the prediction of the
potential for photochromic behaviour was made by comparison of the heats of formation calculated by
AM1. Synthetic routes were explored and as a result two azospirooxazines were isolated, one of which
showed interesting and unusual photochromic behaviour. The instability of this compound towards
purification by chromatography was investigated. The decomposition product was isolated and char-
acterized with the aid of single crystal x-ray crystallography. The correlation between predicted and
observed photochromic behaviour is discussed.
Received 13 November 2012
Received in revised form
20 February 2013
Accepted 22 February 2013
Available online 14 March 2013
Keywords:
Photochromism
Spirooxazine
Azo
Ó 2013 Elsevier Ltd. All rights reserved.
Hydrazone
Molecular modelling
Isosbestic point
1. Introduction
due to a ring-opening reaction. This paper describes the molecular
design, using computer-aided molecular modelling techniques,
There is considerable interest in photochromic materials arising
from the many current and potential applications which are
associated with their ability to undergo reversible, light-induced
colour change. Photochromic dyes are used most widely in
ophthalmic sun-screening applications, and are potentially useful
in security printing, optical recording and switching, solar energy
storage, nonlinear optics and biological systems [1e9]. There has
also been considerable recent interest in their application to tex-
tiles in view of the potential for functional and design applications
[10e17]. Spirooxazines constitute a particularly significant chem-
ical class of photochromic dyes, due to their intense colour gen-
eration properties, good fatigue resistance and ease of synthesis.
The existing range of products generally undergo positive photo-
chromism, a light-induced transition from colourless to coloured
synthesis and evaluation of the photochromic performance of
spirooxazine dyes also containing a permanent azo chromophore,
which exists in the hydrazone form, aimed at producing materials
which undergo a photochromic change from one colour to another.
It was anticipated that the observed colour changes, determined
by the differences in the visible spectra of the ring-closed and
ring-opened forms, would be significantly different from those
given by physical mixtures of permanently coloured dyes and
photochromic dyes, thus providing the potential for niche appli-
cations, for example in anti-counterfeit security printing and brand
protection.
2. Experimental
2.1. Instrumental methods
Melting points were determined as peak tꢁe1mperatures using a
Mettler DSC 12E at a heating rate of 10 ꢀC min from 30 to 400 ꢀC.
Fourier Transform Infrared spectra were recorded as KBr discs with
* Corresponding author. School of Textiles and Design, Heriot-Watt University,
Scottish Borders Campus, Galashiels TD1 3HF, Scotland, UK.
E-mail address: r.m.christie@hw.ac.uk (R.M. Christie).
0143-7208/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.dyepig.2013.02.017

264
R.M. Christie et al. / Dyes and Pigments 98 (2013) 263e272
5
4
5
4
H3C
H3C
6
6
N
O
O
N
O
2'
2'
NO2
7
7
N
N
N
CH3
CH3
10'
7'
CH3
CH3
10'
3''
2''
4
5
9'
9'
8'
H3C
6
N
O
5'
2'
O
HN
7
6'
N
N
N
CH3
CH3
NH
HN
O
2''
3''
2''
3''
8'
5'
7'
6'
NO2
NO2
1
2
3
Fig. 1. Structures of azospiroxazines (illustrated in ketohydrazone forms) 1e3, with numbering scheme for spectral interpretation.
a Nicolet Protégé 460 Spectrophotometer. 1H NMR spectra were
recorded on a Bruker AC 200 instrument at 200 MHz. Mass spectra
were recorded in low resolution Electron Impact mode at 180 ꢀC, on
a Kratos Concept 1S spectrometer. Microanalysis for C, H and N was
carried out on an Exeter Analytical CE440 analyser. UVeVisible
spectra were recorded on a PerkineElmer UVeVIS Lambda 2
spectrophotometer. The photochromism was investigated by irra-
diating solutions contained in silica cells using a Phillips TL 20W/05
UVA bulb, with an emission maximum at 365 nm.
A single crystal of compound 10 suitable for X-ray crystallo-
graphic analysis was obtained by slow cooling of a solution in
ethanol from its boiling point. The single crystal was covered in
Paratone-N and mounted on a cryoloop on a Bruker Nonius X8-
Apex2 CCD diffractometer. Data were collected with MoKa radia-
tion (0.7107 Å) at 100 K, cooled by an Oxford Cryosystems Cryo-
stream. No significant crystal decay was found. Data were collected
for adsorption by psi scans. The structure was solved by direct and
difference Fourier methods and refined by full-matrix least-square
on F2. All non-hydrogen atoms were refined with anisotropic
displacement parameters. Crystallographic computing was per-
formed with SHELXTL programs.
2.2. Synthesis of azospirooxazine 1 (route 1)
2,7-Dihydroxy-1-nitrosonaphthalene (5) was obtained by
nitrosation of 2,7-dihydroxynaphthalene (4) according to a litera-
ture procedure [10]. Spirooxazine 7 was obtained by reaction of
compound 4 with 1,3,3-trimethyl-2-methyleneindoline (Fischer’s
base) (6) according to a literature procedure [18].
2.2.1. Azo coupling of spirooxazine 7 with diazotized p-nitroaniline
p-Nitroaniline (1.04 g, 0.0075 mol) was stirred with concen-
trated hydrochloric acid (7.5 mL). A solution of sodium nitrite
(0.53 g, 0.0075 mol) in water (6.8 mL) was added with stirring over
2 min keeping the temperature below 5 ꢀC with ice cooling. An
excess of nitrous acid was demonstrated by a positive reaction with
starch/KI paper. The mixture was stirred for 20 min at 5 ꢀC. Just
before coupling, a few drops of aqueous sulphamic acid solution
were added to remove the excess of nitrous acid. One fifth of the
total volume was used for the coupling reaction. Spirooxazine 7
(0.34 g, 0.001 mol) was dissolved in ethanol (25 mL). A solution of
sodium hydroxide (0.6 g, 0.15 mol) in water (3.2 mL) was added and
the mixture stirred for 20 min. Anhydrous sodium acetate (2.5 g)
was then added and stirring continued for 30 min. The p-nitro-
benzenediazonium chloride solution was added to the solution of
spirooxazine 7 over 30 min. No diazo excess was observed (H-acid
test). The reaction mixture was stirred overnight. The pH was then
adjusted to 6 by addition of dilute acetic acid. The precipitate was
filtered, washed with water, allowed to dry in air and then recrys-
tallized from ethanol to provide azospirooxazine 1 (0.27 g, 55%) as
Compound 10: C28H25ClN5O5, Mr ¼ 511.53; intense orange
crystal 0.16
ꢂ
0.10
ꢂ
0.08 mm3; triclinic,
a
¼
9.3585(14),
b ¼ 11.1009(17), c ¼ 13.201(2) Å;
a
¼ 111.358(8),
b
¼ 103.385(8),
g
¼
95.229ꢀ;
V
¼
1219.1(3) Å3; space group P1;
Z
¼
2,
Dc ¼ 1.394 Mg mꢁ3, T ¼ 100(2) K;
l
¼ 0.71073 Å (MoKa radiation);
m
¼ 0.098 mmꢁ1. R1 ¼ 0.0507, wR2 ¼ 0.1073 for I > 2
s(I), Goodness-
of-fit on F2 0.973.
O2N
NO2
NO2
O2N
H3C
CH3
CH3
H3C
H3C
CH3
CH3
H3C
N
N
CH3
N
N
N
CH3
H3C
H3C
N
HN
HN
HN
O
HN
N
N
N
N
N
N
O
O
O
O
O
O
O
1a
1b
1c
1d
Fig. 2. The structures of the four possible transoid photomerocyanines from compound 1.

R.M. Christie et al. / Dyes and Pigments 98 (2013) 263e272
265
CH3
CH3
H3C
H3C
H3C
H3C
CH3
CH3
N
CH3
N
N
CH3
H3C
N
N
N
H3C
N
N
N
N
O
O
O
O
O
O
O
O
N
N
NH
NH
NH
NH
NO2
NO2
NO2
NO2
2a
2b
2c
2d
Fig. 3. The structures of the four possible transoid photomerocyanines from compound 2.
00,200
orange-brown crystals. Because of apparent instability of the dye in
solution, carefully-conducted recrystallisation was required. This
involved addition of an optimized volume of boiling ethanol to the
dye and chilling the solution in ice as soon as the dye had dissolved
completely. The purified dye was filtered as soon as a reasonable
quantity of crystals appeared, washed with cold ethanol and dried
in a vacuum dessicator. The compound thus obtained was homo-
geneous by TLC examination. A further recrystallization was used to
provide the analytical sample. The product decomposed without
melting at 220 ꢀC. nmax(KBr)/cmꢁ1: 3435 (NeH st), 2975 (CeH st),
1626 (C]O, weak), 1595, 1572 (aromatic CeC st), 1487, 1334 (NO2
st),1160, 977 (CspiroeO st), 845, 745, def); m/z (EI) (relative intensity,
%): 493 (Mþ, 5), 356 (38), 340 (39), 158 (56), 138 (79), 108 (4), 83
(100), 65 (57), 48 (92); dH ppm (200 MHz, CDCl3) 1.26 (6H, s, (CH3)2),
J3
¼ 9.0 Hz). The UVeVisible spectrum in dichloromethane gave
lmax values of 481 nm, 355 nm and 343 nm with respective molar
extinction coefficients of 2.42 ꢂ 104 Mꢁ1 cmꢁ1, 1.42 ꢂ 104 Mꢁ1 cmꢁ1
and 1.58 ꢂ 104 Mꢁ1 cmꢁ1
In attempts to purify compound 1 (0.27 g, 0.00054 mol) in a
separate experiment, the product was chromatographed on a col-
umn of silica using dichloromethane initially and subsequently
DCM/0.5% acetone. Amide 10 (0.17 g, 33%) was obtained from the
eluates as orange crystals (from ethanol). The product decomposed
without melting at 287 ꢀC. nmax(KBr)/cmꢁ1: 3436 (NeH, st), 1698
(C]O st), 1594, 1490 (Aromatic CeC st), 1336 (NO2 st), 910 (CspiroeO
st), 746; m/z (EI) (relative intensity, %): 511 (Mþ, 7), 509 (79), 481 (15),
372 (40), 359 (18), 344 (11), 331 (4), 254 (6.5),158 (100),144 (45)); dH
ppm (200 MHz, CDCl3); 1.59 (6H, s, (CH3)2), 2.51 (3H, s, NCH3), 3.63
2.86 (3H, s, NCH3), 6.58 (1H, d, 70-H, J7 ,8 ¼ 9.4 Hz), 6.63 (1H, d, 4-H,
(1H, s, CeH), 6.50 (1H, d, 70-H, J7 ,8 ¼ 9.6 Hz), 6.61 (1H, 4-H, m),
0
0
0
0
J4,5 ¼ 8.0 Hz), 6.94 (1 H, m, 5-H), 7.20-7.45 (4H, 6-H, 7-H,50-H, 60-H,
6.70(1H, m, 6-H), 7.06 (1H, m, 80-H), 7.18 (1H, d, 7-H, J7,6 ¼ 8.3 Hz),
m, overlapping signals), 7.58 (1H, d, 80-H, J8 ,7 ¼ 9.8 Hz), 7.71 (2H, d,
7.38 (1H,d, 50-H, J5 ,6 ¼ 7.9 Hz), 7.61 (1 H, m, 5-H), 7.65 (1H,d, 60-H,
0
0
0
0
200-H, J2
¼ 9.2 Hz), 7.88 (1H, 20-H, s), 8.31 (2H, d, 300-H,
J6 ,5 ¼ 8.2 Hz), 7.86 (d, 2H, 200-H, J2
¼ 9.1 Hz), 8.43 (2H,d, 300-H,
0
0
00,300
00,300
CH3
CH3
H3C
H3C
H3C
H3C
N
N
N
CH3
N
N
N
O
CH3
CH3
H3C
H3C
CH3
N
N
O
O
O
O
O
O
O
N
N
N
N
HN
R
HN
R
HN
HN
R
R
3a
3b
3c
3d
Fig. 4. The structures of the four possible transoid photomerocyanines from compound 3.

266
R.M. Christie et al. / Dyes and Pigments 98 (2013) 263e272
Table 1
00,200
J3
¼ 9.1 Hz), 9.80 (s, 1H, OH), 11.65 (br s, 1H, NH), 12.50 (br s, 1H,
Steric energies (kcal molꢁ1) from MM2 calculations for azospirooxazines 1e3 and
their ring-opened merocyanine isomers aed.
NH). Found C, 65.8; H, 4.8; N, 13.2. C28H25N5O5 requires C, 65.7; N,
4.8; N, 13.6%.The UVevisible spectrum in dichloromethane showed
lmax values of 490 nm and 338 nm with respective molar extinction
1
2
3
coefficients of 2.36 ꢂ 104 Mꢁ1 cmꢁ1 and 1.33 ꢂ 104 Mꢁ1 cmꢁ1
.
Closed form
ꢁ1.65
3.42
4.47
3.76
4.65
ꢁ2.76
2.14
ꢁ2.32
2.53
a
b
c
ꢁ0.44
ꢁ0.05
2.3. Synthesis of azospirooxazine 1 (route 2)
3.39
3.93
d
5.17
5.17
P-Nitroaniline (0.276 g, 0.002 mol) was diazotised as described
for Route 1. 2,7-dihydroxy-1-nitrosonaphthalene (5) (0.39 g,
0.002 mol) was dissolved in a solution of sodium hydroxide (0.66 g,
0.016 mol) in water (28 mL) to which anhydrous sodium acetate
(2.76 g) was added as in Route 1. P-Nitrobenzenediazonium chlo-
ride was added to the solution of compound 5 over 30 min. No
diazo excess was observed (H-acid test). The reaction mixture was
stirred overnight and the pH was then adjusted to 6 by addition of
dilute acetic acid. After filtration, washing with water and drying at
60 ꢀC, compound 9 (0.59 g, 85%) was obtained as a highly insoluble
brown powder. Mp 161 ꢀC. nmax(KBr)/cmꢁ1 3431 (N-H, st), 3167 (Ce
Table 2
Heats of formation (kcal molꢁ1) from AM1 calculations for azospirooxazines 1e3
and their ring-opened merocyanine isomers aed.
1
2
3
Closed form
126.83
137.33
140.45
142.79
140.14
126.99
123.20
125.75
126.62
123.78
124.51
135.92
138.70
141.71
137.10
a
b
c
d
HO
OH
4
NO2
NO2
NaNO2 / HCl
+N2
Route 3
NO
HN
N
HO
OH
O
OH
5
8
CH3
H3C
NaNO2 / HCl
NO2
CH2
Route 1
N
Route 2
+N2
NO2
CH3
6
H3C
N
HN
NO
N
N
CH3
CH3
HO
O
O
OH
9
7
CH3
H3C
NO2
CH2
N
1
+N2
CH3
6
Scheme 1. Synthetic routes to azospirooxazine 1.

R.M. Christie et al. / Dyes and Pigments 98 (2013) 263e272
267
H st), 1636 (C]O), 1576, 1437 (Aromatic CeC st), 1343 (NO2 st), 749
(AreH); dH ppm (200 MHz, DMSO-d6); 6.24 (1H, d, 5-H,
J5,6 ¼ 9.5 Hz), 6.77 (1H,d, 3-H, J3,4 ¼ 8.3 Hz), 7.32 (1H, d, 4-H,
J4,3 ¼ 8.2 Hz), 7.86 (1H,d, 6-H, J6,5 ¼ 9.6 Hz), 7.92 (2H,d, H oe to NH,
J ¼ 9.0 Hz), 8.71 (1H, br s, OH), 8.23 (2H,d, H oe to NO2, J ¼ 8.7 Hz).
The compound was unstable to recrystallisation and so was used
subsequently without further purification.
1,3,3-trimethyl-2-methyleneindoline (Fischer’s base) (6) (0.19 g,
0.0011 mol) in ethanol (0.65 mL) was added over 30 min to a so-
lution of nitrosoazo compound 9 (0.338 g, 0.001 mol) in ethanol
(3.8 mL) and the mixture heated under reflux for 2 h. The reaction
mixture was then rotary evaporated to approximately one third of
the original volume and allowed to stand overnight. Filtration,
washing with cold ethanol and drying at 60oCprovided azospir-
ooxazine 1 (0.06 g, 12%) (identical FTIR and NMR spectrum
compared with the product obtained by route 1).
2.4. Attempted synthesis of azospirooxazine 1 (route 3)
P-nitroaniline (3.04 g, 0.022 mol) was diazotised as described for
Route 1. 2,7-dihydroxynaphthalene (4) (3.88 g, 0.024 mol) was
dissolved in a solution of sodium hydroxide (5.48 g) in water
(200 mL). The solution was stirred for 30 min after the addition of
anhydrous sodium acetate (22.85 g). The diazotized p-nitroaniline
solution was added dropwise to the coupling component solution
over 30 min. The final pH was 4.2. The solution was stirred for
75 min, heated to 80e90 ꢀC, filtered, washed with hot water, hot
ethanol and then dried at 60 ꢀC. Azo compound 8 (5.53 g, 75%) was
obtained as an insoluble red powder. Mp. 302 ꢀC. nmax(KBr)/cmꢁ1
:
3404 (NeH, st), 3112 (CeH st), 1653 (C]O), 1590, 1497 (Aromatic
CeC st), 1375 (NO2 st), 748(AreH); found C, 61.9; H, 3.4; N, 13.6;
C10H7NO3 requires C, 62.1; H, 3.6; N,13.6; dH ppm (200 MHz, DMSO-
d6); 6.40 (1H,d, 5-H, J5,6 ¼ 9.6 Hz), 6.88 (1H, d, 3-H, J3,4 ¼ 8.2 Hz),
7.70e7.80 (2H, m, 1-H and 6-H, overlapping signals), 7.50 (1H, d, 4-
H, J4,3 ¼ 8.3 Hz), 7.86 (2H,d, H oe to NH J ¼ 9.2 Hz), 8.31 (2H,d, H oe
to NO2, J ¼ 9.1 Hz). Found C, 61.9; H, 3.4; N, 13.6. C16H11N3O4 re-
quires C, 62.1; N, 3.6; N, 13.6%
Fig. 6. The single crystal x-ray structure of amide 10, with numbering scheme,
showing intramolecular hydrogen bonding.
H st), 1610 (C]O st, weak), 1599, 1572, 1493, 1334 (NO2 st), 1299,
1150, 930(CspiroeO st), 760, 747; m/z (EI) (relative intensity, %): 493
(Mþ, 80), 477 (71), 369 (30), 353 (72), 343 (31), 309 (13), 160 (100),
138 (90), 69 (64); dH ppm (200 MHz, CDCl3) 1.30 (3H, s, CeCH3),1.42
(1H, s, CeCH3), 2.79 (3H, s, NCH3), 6.52 (1H, d, 4-H, J4,5 ¼ 7.7 Hz),
6.89 (1H, m, 5-H), 7.06 (1H, dd, 7-H, J7,6 ¼ 7.79), 7.20 (1H, m, 6-H),
7.45-7.55 (4H, m, 70-H, 80-H, 90-H, 100-H), 7.60 (2H, d, H oe to NH,
J10,11 ¼ 9.16 Hz), 7.97 (s, 1H, 20-H), 8.30 (2H, d, H oe to NO2,
J12,13 ¼ 9.0 Hz), 15.8 (1H, br s, NH). The UVeVisible spectrum in
dichloromethane showed lmax values of 495 nm and 345 nm with
respective molar extinction coefficients of 2.9 ꢂ 104 Mꢁ1 cmꢁ1 and
1.3 ꢂ 104 Mꢁ1 cmꢁ1. Since the compound was not obtained in
analytically pure form, these molar extinction coefficients should
be regarded as approximate.
Attempted nitrosation of compound 8 using a range of standard
procedures led only to recovery of starting material.
2.5. Synthesis of azospirooxazine 2
Spirooxazine 13 was prepared from 2,3-dihydroxynaphthalene
(11), via 2,3-dihydroxy-1-nitrosonaphthalene (12) according to
literature procedures [18].
Diazotisation of p-nitroaniline and coupling with spirooxazine
13 was carried out as for the synthesis of dye 1 by route 1. Azo-
spirooxazine 2 (0.27 g, 77%) was provided as brown crystals ob-
tained by recrystallisation from ethanol (x 2). The compound
decomposed without melting at 196 ꢀC. nmax(KBr)/cmꢁ1: 3436 (Ne
2.6. Attempted synthesis of azospirooxazine 3
Spirooxazine 16 was prepared from 2,6-dihydroxynaphthalene
(14), via 2,6-dihydroxy-1-nitrosonaphthalene (15) according to
literature procedures [18].
Table 3
Hydrogen bonds for compound 10 [Å and ꢀ].
D-H.A
d(D-H)
d(H.A)
d(D.A)
<(DHA)
N(1)-H(1N).N(9)
N(1)-H(1N).N(13)
N(10)-H(10N).O(8)
O(2)-H(2O).O(11)
C(6)–H(6)..O(26)i
0.865(18)
0.865(18)
0.93(2)
0.96(3)
0.95
2.025(17)
2.263(17)
1.80(2)
1.63(3)
2.47
2.711(2)
2.734(2)
2.569(2)
2.5713(18)
3.396(2)
3.288(2)
3.206(2)
135.5(15)
114.2(14)
139.1(18)
163(2)
165
C(13)–H(13C)..O(8)ii
C(25)–H(25)..O(2)iii
0.98
0.95
2.49
2.52
139
129
Symmetry transformations used to generate equivalent atoms: i: x, ꢁ1 þ y, ꢁ1 þ z;
ii: 1 ꢁ x, 2 ꢁ y, ꢁz; iii: 1 ꢁ x, 2 ꢁ y, ꢁz.
Fig. 5. Structure of amide 10.

268
R.M. Christie et al. / Dyes and Pigments 98 (2013) 263e272
An attempt was made to synthesise azospirooxazine 3 from
molecules are likely to exist in the ketohydrazone rather than the
hydroxyazo tautomeric form and in a conformation which favours
intramolecular hydrogen bonding as illustrated in Fig. 1 [20]. Spi-
rooxazines owe their photochromic properties to a light-induced
ring-opening to form a photomerocyanine, which reverts ther-
mally to the ring-closed form when the light source is removed.
There are eight possible isomers of the photomerocyanines, four of
which may be considered as cisoid and four as transoid. The cisoid
isomers are likely to be highly unstable due to steric constraints so
that, in common with assumptions made previously, the investi-
gation centred on the transoid isomers aed (Figs. 2e4).
compound 16 using a procedure as for the synthesis of dye 1 by
route 1. A low yield of a crude brown product was obtained. At-
tempts to purify by column chromatography and recrystallization
showed that compound is highly unstable, with a tendency to
generate new impurities during the purification process.
3. Results and discussion
Azospirooxazines 1e3 (Fig. 1) were selected as target molecules
based to a certain extent on an assessment of synthetic feasibility.
Prior to attempted synthesis, the molecules were subjected to
computer-aided molecular modelling to provide a prediction of
features of their photochromism based on a methodology previ-
ously devised and verified in these laboratories using the facilities
provided by the CAChe system [19]. It was assumed, based on
established knowledge of the structures of azonaphthols, that the
The molecular geometries of the azospirooxazines 1e3 and each
of the isomers of the probable photomerocyanines were calculated
using standard augmented MM2 within CaChe, with
a fine
convergence limit of 1eꢁ5 used in the minimisations, and this
approach was also used to provide steric energies (Table 1). Heats of
formation were calculated using AM1 (Table 2).
NO2
H+
H3C
H2O:
N
O
1
HN
+
N
HN
CH3
CH3
O
-H+
NO2
NO2
H3C
+
H3C
H
N
H
HO
HN
H+
N
O
HO
HN
HN
HN
N
N
CH3
CH3
CH3
CH3
O
OH
O
-H+
NO2
H3C
tautomerism
N
HO
HN
10
HN
N
CH3
CH3
OH
O
Scheme 2. Mechanism proposed for the conversion of azospirooxazine 1 to amide 10.

R.M. Christie et al. / Dyes and Pigments 98 (2013) 263e272
269
Fig. 7. The optimized (MM2) geometric structures of compound 1 and its ring-opened forms.
Both MM2 and AM1 calculations predict that compounds 1 and
3 would be expected to exist normally in the ring-closed form since,
in each case, it is of lower energy than all of the isomers of the ring-
opened forms. This suggests that these dyes might have the po-
tential to show photochromism, by absorption of UV light to
convert to the higher energy states, with thermal reversion to the
more stable ring-closed form when the light source is removed. The
prediction is ambiguous in the case of compound 2. AM1 calcula-
tions suggest that the ring-opened forms are marginally more
stable than the ring-closed form, predicting the possibility of a
permanent merocyanine structure, while MM2 calculations sug-
gest the opposite. However, the calculated values for the various
species are very close in these cases.
For compound 1, three complementary synthetic routes were
attempted as illustrated in Scheme 1. Each route involved three
stages: azo coupling of a dihydroxynaphthalene derivative with
diazotized p-nitroaniline, nitrosation and reaction with Fischer’s
base (6). The three routes differed in the sequence in which the
reactions were carried out. Azospirooxazine 1 was successfully
synthesized by both routes 1 and 2. Isolation of compound 1
initially proved problematic. The compound was unstable to a
range of chromatographic conditions, and to a certain extent
thermally during recrystallisation. Ultimately an optimized syn-
thetic procedure was developed, on the basis of repeated trials,
leading to a product which TLC examination indicated was rela-
tively free from impurities, and could be purified to a homogeneous
product by a carefully-conducted recrystallisation process, as
described in the experimental section. Route 1 proved to be the
best. Under optimized conditions, azo coupling of spirooxazine 7
with p-nitrobenzenediazonium chloride in aqueous ethanol gave
the orange-brown azospirooxazine 1 in 55% yield. Route 2 gave a
much lower overall yield and presented more difficulties at the
purification stage. Route 3 failed at the nitrosation stage, probably
because of the extreme insolubility of azo compound 8 in the acidic
medium. The existence of compound 1 as the ring-closed spi-
rooxazine was demonstrated by 1H NMR spectroscopy, for example
showing the singlet characteristic of the oxazine proton (20-H) at
d
7.88 ppm.
0.4
Before UV exposure
5 min. UV
10 min. UV
20 min. UV
40 min. UV
60 min. UV
0.2
0
300
450
350
400
500
550
600
650
700
Wavelength, nm
Fig. 8. UV/visible spectra of solutions of azospiroxazine 1 in ethanol with UV exposure
Fig. 9. UV/visible spectra of a solution of azospiroxazine 1 in ethanol following two
with time up to 80 min.
cycles of UV exposure for 20 min, with 24 h in the dark between exposures.

270
R.M. Christie et al. / Dyes and Pigments 98 (2013) 263e272
It had been noted during our attempts at optimization of the
Fig. 7 illustrates the structures of compound 1 and the possible
photomerocyanines 1ae1d, as modelled by MM2 calculations.
Compound 1 shows a molecular geometry with the p-nitro-
phenylhydrazone system more or less in the same plane as the
naphthoxazine system, which is orthogonal to the plane of the
indoline system as a consequence of the spiro arrangement. As
discussed earlier, the MM2 and AM1 calculations suggest that
compound 1 has the potential to exhibit photochromism. Indeed,
we have found that azospiroxazine 1 shows interesting and un-
usual photochromic properties, converting in ethanolic solution by
UV irradiation from an orange colour to a neutral grey, progres-
sively and slowly in comparison with commercial spirooxazine-
based photochromic dyes.
synthesis that compound 1 was unstable to TLC on silica and to
column chromatography, converting to an orange compound. By
deliberately exposing compound 1 to preparative silica column
chromatography, the orange product was obtained in 33% yield.
This compound was demonstrated to be the secondary amide 10
(Fig. 5) on the basis of spectroscopic analysis confirmed by a single
crystal x-ray structure determined from crystals grown from
ethanol. The crystal structure (Fig. 6) demonstrates that the com-
pound exists in the ketohydrazone form in which there is signifi-
cant intramolecular hydrogen bonding in two 6-membered rings
(between hydrazone NeH and ketone C]O and between amide Ne
H and hydrazone C]N) and in a 7-membered ring (between OeH
and amide C]O). Table 3 shows relevant interatomic distances and
angles. With regard to intermolecular interactions, there is poten-
Fig. 8 shows the UV/visible spectra of azospiroxazine 1 during
irradiation with UV light with exposure times up to 60 min. The
ring-closed compound shows a single visible absorption band
tial
p
ꢁstacking where the distance between the centroids of the
two symmetry equivalent rings C23eC28 related by a centre of
inversion is 3.5580(12) Å. The closest intermolecular contacts are
three CH..O interactions where the C..O distances are in the range
3.206(2)e3.396(2) Å.
(
lmax ¼ 481 nm), associated mainly with the azo (hydrazone)
group, which is responsible for its orange colour. The spectrum of
the photoproduct after 60 min of UV exposure shows two ab-
sorption bands in the visible region, lmax ¼ 487 nm, close to the
original absorption band but with lower intensity and a change of
shape, and a broad longer wavelength band, lmax ¼ 571 nm. The
absorption throughout the visible region accounts for the observed
grey colour. There are also differences in the UV spectral region. An
important observation in Fig. 8 is the definitive presence of iso-
sbestic points at 353, 392 and 528 nm. Isosbestic points are loca-
tions in absorption spectra at which the species in solution absorb
with equal intensity at particular wavelengths, and their presence
provides evidence that only two principal species are present in the
solution during the conversion [21e24]. This feature is consistent
A plausible mechanism for the formation of 10 is proposed in
Scheme 2. This proposition suggests an initial acid-catalysed (by
silica) nucleophilic addition of water, presumably from within the
silica, followed by ring-opening by cleavage of the weak CspiroeO
bond. This feature may account for observations of the instability of
spirooxazines towards aqueous acidic conditions and this reaction
may provide an insight into possible hydrolytic degradation
mechanisms [12]. The ability to isolate a product, compound 10 in
this particular case, is probably due to the stability provided by the
extensive intramolecular hydrogen bonding.
Fig. 10. The optimized (MM2) geometric structures of compound 2 and its ring-opened forms.

R.M. Christie et al. / Dyes and Pigments 98 (2013) 263e272
271
with the photochromic behaviour of azospiroxazine 1 involving a
photochemical conversion of the ring-closed form to another spe-
cies, the value of the absorption maximum of which (571 nm) is
consistent with that anticipated for a photomerocyanine [20]. The
spectrum of the unexposed solution passes through the isosbestic
points at 353 and 528 nm, but does not pass so accurately through
the isosbestic point at 392 nm, indicating that the mechanism may
not be as simple as this. It is observed that the absorption spectrum
after UV exposure for longer than 60 min begins to deviate from
isosbestic behaviour, an effect which becomes more pronounced as
exposure times are increased. This observation indicates the for-
mation of new species, possibly due to photodegradation as a result
of prolonged exposure to UV light. A further possibility which may
contribute towards this observation is chemical degradation in
ethanol, especially in view of the hydrolytic instability of dye 1.
When the UV light source is removed after exposure, there is a
slow colour change when the solution is stored in the dark,
although it does not revert to the original orange colour. The
spectral change when a sample is irradiated for 20 min and then is
allowed to fade in the dark for 24 h at normal temperatures is
shown in Figs. 9 and 10. The spectrum shows a decrease in the
longer wavelength absorption, as might be expected from a ther-
mal reversal by ring closure of the merocyanine form, although
there is little change in the lower wavelength absorption. A second
irradiation of this sample shows some, but rather limited, residual
photochromic behaviour with an increase in the longer wavelength
absorption and a decrease in the lower wavelength band intensity
(Fig. 9). The level of photochromism due to the second irradiation is
observed to reduce further with longer initial UV exposure times.
This behaviour is consistent with partial reversible photochromism,
but accompanied by degradation processes in solution leading to
non-photochromic products. Additional investigation of the
photochemical behaviour of the dye in solution has demonstrated
its complexity, for example that it is strongly solvent-dependent.
OH
OH
OH
HO
11
14
NaNO2/HCl
NO
NO
OH
OH
HO
OH
15
12
CH3
H3C
CH2
N
CH3
6
H3C
H3C
N
N
N
N
CH3
CH3
CH3
CH3
O
O
OH
HO
16
13
+N2
NO2
3
2
Scheme 3. Synthetic routes to azospiroxazines 2 and 3.

272
R.M. Christie et al. / Dyes and Pigments 98 (2013) 263e272
Details of an extensive study will form the basis of a future
publication.
degradation leading to non-photochromic products. Dye 1 was
converted on a silica chromatography column to a secondary amide
(10), whose structure was confirmed by single crystal x-ray crys-
tallography. A mechanism for the conversion is proposed, which
may provide an insight into the reasons underlying the instability
of spirooxazines towards aqueous acidic conditions.
Steric energies from MM2 calculations and heats of formation
values obtained from AM1 calculations (Tables 1 and 2) suggest
that isomer 1a is the lowest energy ring-opened form and thus is
likely to be the dominant species in the photomerocyanine which is
observed experimentally when compound 1 is irradiated with UV
light, on the basis of the UV/visible spectra in Fig. 8. Inspection of
the 3D structures calculated for 1ae1d indicate that there is
considerable steric congestion in each of the isomers forcing sig-
nificant non-planarity, but that this is minimised in isomer 1a
which deviates least from planarity (Fig. 7).
Since route 1 proved to be optimal for synthesis of compound 1,
attempts were made to synthesise compounds 2 and 3 following
analogous procedures. Azospirooxazine 2 was obtained as a brown
solid in low yield starting from 2,3-dihydroxynaphthalene (11) by
the route illustrated in Scheme 3. Isolation of dye 2 in analytically
pure form proved problematic because of instability towards the
purification process. In this case, the prediction of photochromic
properties based on MM2 and AM1 calculations was ambiguous
(Tables 1 and 2). However, the spectral data, especially by com-
parison of its 1H NMR spectrum with that of the isomeric dye 1,
demonstrate that dye 2 exists as the ring-closed form, for example
showing the singlet characteristic of the oxazine proton (20-H) at
Acknowledgements
We thank Dr Mark Heron, Department of Chemistry, University
of Huddersfield, for helpful and informed discussion. We thank
the Worshipful Company of Dyers of the City of London and the
University of Damascus, Syria, for financial support, in the form of
the award of PhD studentships (to AR and MA respectively).
References
[1] Brown GH, editor. Photochromism. New York: John Wiley & Sons; 1971.
[2] El’tsov AV. Organic photochromes. New York: Plenum Publishing Corporation;
1990.
[3] Durr H, Bouas-Laurent H, editors. Photochromism: molecules and systems.
New York: Elsevier; 1990.
[4] McArdle CB, editor. Applied photochromic polymer systems. Glasgow and
London: Blackie; 1992.
[5] Crano JC, Guglielmetti RJ, editors. Organic photochromic and thermochromic
compounds. New York: Plenum Press; 1999.
[6] Berkovic G, Krongauz V, Weiss V. Spiropyrans and spirooxazines for memories
and switches. Chem Rev 2000;100:1741e54.
d
7.97 ppm, whereas the merocyanine would have given the cor-
responding proton at much higher field. In this case (Fig. 10), there
is clearly much less steric interaction in the ring-opened forms 2a-
2d than in 1a-1d and the isomers are calculated to be nearly
coplanar. We have investigated UV irradiation of solutions of azo-
spirooxazine 2, but have not been able to demonstrate photo-
chromic properties. However, photochromism can be dependent on
a range factors, including the nature of the solvent, the temperature
and the irradiation conditions and thus it cannot be stated with
certainty that the dye is non-photochromic.
[7] Tyson DS, Bignozzi CA, Castellano FN. Metal-organic approach to binary
optical memory. J Am Chem Soc 2002;124:4562e3.
[8] Willner I. Photoswitchable biomaterials: en route to optobioelectronic
systems. Acc Chem Res 1997;30:347e56.
[9] Towns AD. Colour is the mother of invention. Color Technol 2011;127 [iii].
[10] Billah SMR, Christie RM, Shamey R. Direct coloration of textiles with
photochromic dyes. Part 1: application of spiroindolinonaphthoxazines as
disperse dyes to polyester, nylon and acrylic fabrics. Color Technol 2008;124:
223e8.
[11] Billah SMR, Christie RM, Morgan KM. Direct coloration of textiles with
photochromic dyes. Part 2: the effect of solvents on the colour change of
photochromic textiles. Color Technol 2008;124:229e33.
Attempts to synthesise dye 3 from 2,6-dihydroxynaphthalene
(14) as illustrated in Scheme 3 gave, in the final stage, a low yield of
a crude brown product which proved highly unstable towards at-
tempts at purification both by recrystallisation and using chroma-
tography, processes which invariably gave rise to the formation of a
range of coloured impurities so that the compound could not be
obtained with sufficient purity for rigorous characterization.
Nevertheless, there is reasonable evidence from FTIR and 1H NMR
spectra, in comparison with those obtained for dyes 1 and 2, and
from TLC behaviour, that dye 3 in its ring-closed form is the major
component of the crude product. Characteristic singlets in the 1H
[12] Billah SMR, Christie RM, Shamey R. Direct coloration of textiles with photo-
chromic dyes. Part 3: dyeing of wool with photochromic acid dyes. Color
Technol 2012;128:488e92.
[13] Little AF, Christie RM. Textile applications of photochromic dyes. Part 1:
establishment of a methodology for evaluation of photochromic textiles using
traditional colour measurement instrumentation. Color Technol 2010;126:
157e63.
[14] Little AF, Christie RM. Textile applications of photochromic dyes. Part 2:
factors affecting the photocoloration of textiles screen-printed with com-
mercial photochromic dyes. Color Technol 2010;126:164e70.
[15] Little AF, Christie RM. Textile applications of photochromic dyes. Part 3:
factors affecting the technical performance of textiles screen-printed with
commercial photochromic dyes. Color Technol 2011;127:275e81.
[16] Aldib M, Christie RM. Textile applications of photochromic dyes. Part 4:
application of commercial photochromic dyes as disperse dyes to polyester by
exhaust dyeing. Color Technol 2011;127:282e7.
[17] Aldib M, Christie RM. Textile applications of photochromic dyes: part 5.
Application of commercial photochromic dyes to polyester fabric by a solvent-
based dyeing method, submitted for publication in Color Technol.
[18] Osterby B, McKelvey RD, Hill L. Photochromic sunglasses: a patent-based
advanced organic project and demonstration. J Chem Educ 1991;68:424e5.
[19] Christie RM, Chi LJ, Spark RA, Morgan KM, Boyd ASF, Lycka A. The application
of molecular modeling techniques in the prediction of the photochromic
NMR spectrum at
were observed together with the signals due to the AB system
characteristic of the p-nitroaniline derived aromatic ring system.
d
1.35 (CeCH3), 2.70 (NCH3) and 7.82 ppm (20-H)
4. Conclusions
Spirooxazines containing an azo chromophore, which exists in
the hydrazone form, have been synthesized in three stages from
dihydroxynaphthalenes. Our previously-established methodology
involving MM2 and AM1 calculations was used to predict aspects of
photochromic behaviour in the selected synthetic target molecules.
behaviour of spiroindolinonaphthoxazines.
J Photochem Photobiol A
2005;169:37e45.
[20] Christie RM, Chang CH, Huang H-Y-, Vincent M. Colour and constitution
relationships in organic pigments: part 6 azonaphtharylamide pigments. Surf
Coat Int Pt B-c 2006;89:77e85.
[21] Lower SK. Thermal physics (and some chemistry) of the atmosphere. J Chem
Educ 1998;75:837e40.
Azospirooxazine
1 shows interesting, unusual photochromic
properties. The dye converts slowly by UV irradiation in ethanolic
solution from an orange colour to grey, the spectrum being
consistent with the formation of a photomerocyanine, in a process
showing isosbestic spectral behaviour. However, the reverse ther-
mal reaction appears to be incomplete and is complicated by
[22] Trogler WC. The environmental chemistry of trace atmospheric gases. J Chem
Educ 1995;72:973e6.
[23] Harris DC, editor. Quantitative chemical analysis. 7th ed. New York: W. H.
Freeman; 2007.
[24] Geissler PL. Temperature dependence of inhomogeneous broadening: on the
meaning of isosbestic points. J Am Chem Soc 2005;127:14930e5.