Effect of alkali on methylene blue (C.I. Basic Blue 9) and other thiazine dyes

Article abstract of DOI:10.1016/j.dyepig.2010.05.015

A detailed study of the action of alkali on methylene blue (C.I. Basic Blue 9) and other thiazine dyes was carried out through a combination of UV/visible spectroscopy, thin layer chromatography, mass and NMR spectrometry and computational methods. In 0.1 M aq alkali solution, methylene blue forms a highly coloured, lipophilic species that is mainly Bernthsen's methylene violet i.e. a hydrolysis decomposition product, this being contrary to the report of a red N-hydroxy methylene blue adduct. The nature of the heterocyclic nitrogen atom in C.I. Basic Blue 9 is discussed and it is concluded there is no basis for the proposal of nucleophile addition at this site of the dye. In contrast, other thiazine dyes are deprotonated by alkali to form their neutral, highly coloured, lipophilic conjugate base forms.

Full text of DOI:10.1016/j.dyepig.2010.05.015

Dyes and Pigments 88 (2011) 149e155
Contents lists available at ScienceDirect
Dyes and Pigments
journal homepage: www.elsevier.com/locate/dyepig
Effect of alkali on methylene blue (C.I. Basic Blue 9) and other thiazine dyes
,
Andrew Mills a, David Hazafy a, John Parkinson a, Tell Tuttle a, Michael G. Hutchings b
*
a Department of Pure & Applied Chemistry, University of Strathclyde, Glasgow G1 1XL, UK
b School of Chemistry, University of Manchester, Manchester M13 9PL, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
A detailed study of the action of alkali on methylene blue (C.I. Basic Blue 9) and other thiazine dyes was
carried out through a combination of UV/visible spectroscopy, thin layer chromatography, mass and NMR
spectrometry and computational methods. In 0.1 M aq alkali solution, methylene blue forms a highly
coloured, lipophilic species that is mainly Bernthsen’s methylene violet i.e. a hydrolysis decomposition
product, this being contrary to the report of a red N-hydroxy methylene blue adduct. The nature of the
heterocyclic nitrogen atom in C.I. Basic Blue 9 is discussed and it is concluded there is no basis for the
proposal of nucleophile addition at this site of the dye. In contrast, other thiazine dyes are deprotonated
by alkali to form their neutral, highly coloured, lipophilic conjugate base forms.
Received 1 March 2010
Received in revised form
12 May 2010
Accepted 25 May 2010
Available online 11 June 2010
Keywords:
Methylene blue
Methylene violet
C.I. Basic Blue 9
Thiazine
Ó 2010 Elsevier Ltd. All rights reserved.
Hydrolysis
DFT
1. Introduction
also be anticipated to be colourless, thereby calling into doubt the
structural assignment of the red material isolated from alkali
C.I. Basic Blue 9 (Methylene Blue (MB; 1; Table 1)) is a classical,
basic dye, originally synthesized by Heinrich Caro, that has been
commercially available since its first production by BASF in 1876
[1]. It enjoys manifold, widespread uses such as a dye for hair,
leather and cellulosic fibres, redox indicator, ISO test pollutant in
semiconductor photocatalysis, photosensitizer for singlet oxygen
generation, antioxidant and antiseptic, stain for fixed and living
tissues, diagnostic agent in renal function tests, antidote to cyanide
and nitrate poisoning and as a treatment for malaria [2,3]. It has
recently been reported to be effective in arresting the progress of
Alzheimer’s and other neurodegenerative diseases [4e6].
Amazingly, this far-reaching and sustained interest shows no
evidence of fading insofar as >4000 MB-related papers were
published in the last 5 years. However, there is still much about this
apparently simple dye molecule that can surprise. For example, it
was recently reported that MB readily forms a red-coloured
treatment of MB, which, in any case, is in conflict with the classical
literature which reports the main product to be the hydrolysed
species, Bernthsen’s methylene violet (MVB; C.I. 52041; 4) [9]. In
view of MB’s continuing and topical applications, the chemistry of
MB and related examples was reviewed; a brief overview of the
results has been published elsewhere as a comment [10] on the
original paper [7], accompanied by a reply [11].
The earlier assignment [7] of structure
2 was based on
a proposal by Plater [12] that the heterocyclic nitrogen of MB is
electron deficient and thus a target for nucleophilic attack.
Although the current authors considered this proposal unlikely,
comments from referees and others induced us to address this
point in more detail in this paper. The objectives of this work were
to present the full evidence for the characterization of the product
obtained from reaction between MB and NaOH, to discuss the
nature of MB’s heterocyclic N atom alongside other related dye
species and to summarize the results of investigations into how
other thiazine-based dyes behave towards alkali.
(
lmax ¼ 525 nm in toluene) lipophilic, non-ionic hydroxy adduct
upon treatment with an equimolar amount of sodium hydroxide, to
which the N-hydroxy structure MB-OH 2 was assigned [7].
Structure 2 is closely related to the leuco form of MB (3) which is
almost colourless [8]. Consequently a molecule of structure 2 would
2. Experimental
2.1. Materials and chemicals
Methylene blue (96þ%, as hydrate, MW. 319.85; MB (1)) was
purchased from Acros. Azure B (80%þ, Mr 305.83; C.I. 52010; AB (5)),
* Corresponding author.
E-mail address: mike.hutchings@manchester.ac.uk (M.G. Hutchings).
0143-7208/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.dyepig.2010.05.015

150
A. Mills et al. / Dyes and Pigments 88 (2011) 149e155
Table 1
Summary of structural, spectral and physical characteristics of MB and other relevant thiazine dyes and derivatives.
c
Structure
No.
Abbreviation, name,
molecular weight
lmax/nm
pKa
Cationa
Neutralb
N
S
MB
+
(1)
Methylene blue
284.4
665 (746)d
Me2N
NMe2
NMe2
OH
N
MB-OH
256.32
520f
[351]
(2)
(3)
Me2N
S
H
N
Leuco-MB
285.4
256g
520h
Me2N
Me2N
S
NMe2
O
N
S
MVB
Methylene violet
256.32
(4)
(5)
(6)
610e
645
628
N
S
AB
Azure B
270.37
+
503
[521]
12.1
11.8
Me2N
NHMe
NH2
N
S
AA
Azure A
256.35
+
500
[501]
Me2N
N
S
AC
Azure C
242.32
+
488
[520]
(7)
(8)
615
602
11.5
11.0
MeHN
NH2
N
S
TH
Thionine
228.29
487
[486]
+
H2N
NH2
a
Absorption maxima of cationic species measured in water.
b
c
Absorption maxima of neutral deprotonated species measured in toluene. Calculated values in square brackets []. See Discussion, Section 3.3.
Data from Bonneau R, Faure J, Joussot-Dubien J. Talanta, 1967; 14:121.
Protonated doubly charged absorption maximum from reference [40].
O-protonated MVB.
Data from reference [7], assigned to MB-OH but actually for MVB.
Data from Obata H. Bull Chem Soc Jpn 1961; 34:1057; Cohn G. Ber 1900; 33:1567.
Data for neutral unprotonated MVB.
d
e
f
g
h
Azure A (96%, Mr 291.80; C.I. 52005; AA (6)), Azure C (40%, Mr 277.77;
C.I. 52002; AC (7)), thionine (85%, Mr 287.34; C.I. 52000; TH (8)) and
methylene violet (Bernthsen), (80%, Mr 256.33; C.I. 52041; MVB (4))
were purchased from Aldrich; all were chloride salts except thionine
acetate. Structures are given in Table 1. MB was further purified by
continual extraction of red impurities from its buffered aqueous
solution by CCl4 until the latter washings were colourless, followed by
crystallization according to a published method [13]. All other
chemicals were used without further purification. Analytical thin
layer chromatography was carried out using silica plates employing
MeOH/CHCl3/HOAc 90:8:2 as mobile phase.
quartz cuvettes in dry solvents of highest available purity. A Per-
kineElmer fluorimeter LS 50 B was used to record fluorescence
spectra.
2.3. Mass spectrum analysis
Two different mass spectrometers were used in this work. The
majority of the experiments were run on an ESI-MS (Thermo-
Finnigan LCQ DUO MS) using either the direct injection port or via
an HPLC system (i.e. LC-MS). In the latter, the HPLC column was
a Phenomenex Gemini, C18, 5
phase was used comprising initially 0.1% formic acid in water,
becoming 0.1% formic acid in acetonitrile, over 15 min at 150 l/min
mm, 50 ꢀ 2.0 mm. A graded mobile
2.2. Spectrophotometric measurements
m
flow rate. An LDI-MS (Shimadzu, AXIM-CFR) was also used to
record the mass spectra of the thiazine dyes as well as the MB base
hydrolysis products. Samples of the dyes were prepared and
All UV/visible spectrophotometric experiments were recorded
on a Varian Cary 50 double beam spectrophotometer, using 1 cm

A. Mills et al. / Dyes and Pigments 88 (2011) 149e155
151
recorded at different pH values (neutral, 10, 12) and the toluene
extract prepared by shaking the dye (0.1 mM) in 0.1 M NaOH with
an equal volume of toluene (vide infra).
Fig. 1 shows photographs of equal volumes of an aqueous MB
(0.1 mM) solution below toluene at time zero and 5 h after the
addition of sufficient alkali to render the aqueous solution pH 13
and mixing via thorough shaking. The observed UV/visible spectra
of the MB aqueous solution at pH 13 and toluene solution before
and 1 h and 5 h after mixing are also shown in Fig. 1. These findings
show that at pH 13 MB is converted to a lipophilic, i.e. toluene-
soluble, red/pink species which is clearly observable after 1h and
for which the reaction is complete after 5 h. It has the same spectral
features as the MB-derived species reported [7] but is it really
a N-hydroxy adduct?
2.4. Computational methods
All structures were optimized in the solvent phase using the
polarisable continuum model [14,15] (PCM) at the B3LYP [16e21]
level of theory with the 6ꢁ311þþG(d,p) basis set [22,23]; no
symmetry constraints were imposed during the optimization of the
dyes. The structures were optimized in two different solvent
(dielectric) environments, water (
3
¼ 78.4) and toluene ( ¼ 2.4). Time
3
dependent density functional theory [24e26] (TD-DFT) single-point
calculations were performed on the optimized structures to obtain
the calculated lmax values. The PCM approach was employed within
the TD-DFTcalculations to model the effect of the respective solvents
on the absorption spectra. All calculations were done within the
Gaussian 03 program [27]. The charge distribution of MB was
determined using the natural bond orbital (NBO) approach [28]. The
calculated values for the absorption lmax of relevant, different thia-
zine dye species in toluene are given in Table 1.
3.2. Properties of “red methylene blue”
A number of simple experiments reveal the red MB produced as
described above using a pH 13 aqueous solution of MB is not
a hydroxy adduct, but rather largely Bernthsen’s methylene violet,
MVB (4) [32]. The free base form of the latter is not very soluble in
water (0.6 mg mLꢁ1, cf. MB 50 mg mLꢁ1), but is soluble in most
common organic solvents, including toluene. Thus, the UV/visible
absorption spectrum of a commercial sample of MVB (4) dissolved
in toluene is identical to that of red MB and both also fluoresce with
2.5. NMR measurements
the maximum of emission at 596 nm (
l
(excitation) ¼ 520 nm in
1H NMR data were acquired using Bruker AVANCE-III and
Avance/DRX NMR spectrometers operating at 600.13 and
500.13 MHz and operating under TopSpin versions 2.0 and 1.3
respectively. Data accumulated for samples solubilized in CCl4 were
acquired in an unlocked mode, with magnetic field homogeneity
adjusted manually using lineshape observation-based shimming.
toluene), in agreement with the values reported in the literature for
MVB fluorescence [33]. The measured UV/visible spectra of MVB (4)
in different organic solvents are very similar if not indistinguishable
from the spectra reported [7] in the same solvents indicating they
are the same. In support of this, the hydrolysis of MB with alkali to
3. Results and discussion
3.1. MB initial experiments: formation of Red MB
The red lipophilic form of MB assigned the MB-OH structure 2
[7], henceforth referred to as ‘red MB’, was reportedly generated by
mixing an ‘aliquot’ of aqueous MB (0.1 mM) with NaOH (0.1 mM, i.e.
pH 10) under toluene [7]. After 1 h standing, the toluene reportedly
developed a red colour due to the extraction of the lipophilic red
MB from the aqueous solution. However, in our hands, upon
reproducing this simple experimental procedure, no red MB was
generated. This was not surprising, given others have noted MB is
‘indefinitely stable’ in aqueous solution at pH 9.5 [29,30]. Indeed,
mixing MB with 0.1 mM alkali and shaking with a water-immiscible
solvent (usually dichloromethane) is a published method for
purifying MB of the less methylated thiazines, such as azure B
(5;usually the most prevalent species), azure A (6) and azure C (7),
which are common impurities in most past commercial samples of
MB [13,31]. It is generally accepted that this purification procedure
is effective because the latter thiazines are readily deprotonated by
the alkali to their neutral, lipophilic orange or red-coloured forms
[31] and MB itself is stable in 0.1 mM alkali. More about this process
will be discussed later.
It is tempting to explain our failure to reproduce the formation of
red MB as being due to the earlier [7] use of a source of MB that was
contaminated with one or more of the thiazines listed in Table 1.
However, the reported absorption spectrum and lmax (526 nm) for
red MB is not that of any of the deprotonated thiazines. Interest-
ingly, it is possible to generate a species, with a near identical
absorption spectrum to that reported for red MB [7], using the same
method, but with 0.1 M, instead of 0.1 mM, NaOH aqueous phase
solution shaken with toluene. The simplest explanation for the
formation of the red dye in the original account would appear to be
the use of a more concentrated alkaline solution than reported [7].
Fig. 1. Top: Photographs of a fresh 10ꢁ4 M MB, 0.1 M NaOH aqueous solution (from left
to right): before, directly after and 5 h after mixing and shaking with an equal volume
of toluene. Bottom: Visible absorption spectrum (measured in 1 cm cuvette) of the
aqueous MB solution (i) before, (ii) directly following, and (iii) 5 h after mixing and
shaking with an equal volume of toluene. The visible absorption spectra of the toluene
solution directly after and 5 h following mixing and shaking with the aqueous solution
are illustrated by lines (iv) and (v).

152
A. Mills et al. / Dyes and Pigments 88 (2011) 149e155
produce MVB (4) has been suggested previously by others [9]
usually based on spectral observations.
(5; Table 1) [34]. As already stated this is a common MB contami-
nant, and can be readily formed from MB by alkali-induced
demethylation [35,36] at pH 11, a value somewhat lower than that
(pH 13) which causes deamination to MVB.
Further confirmation of the identity of red MB was afforded by
the observation that when it is re-extracted from toluene by an
acidic aqueous layer, the latter turns blue with an absorption spec-
trum that is not MB (lmax ¼ 665 nm) but rather that of protonated
MVB (lmax ¼ 580 nm). When the red MB in toluene solution is
spotted onto an acidic silica TLC plate it also turns blue, as noted
earlier [7,12], but tlc development causes the original dye spot to
separate into two blue spots, neither of which is MB namely, one (the
most striking in terms of depth of colour) has the retention time, Rf,
of that of commercial MVB (4) and the other, which is much weaker
in colour, has an Rf value the same as that of azure B (5; see Table 1).
The NMR spectra recorded for MB (in D2O; 500.13 MHz), red MB
and MVB (in CCl4; 600.13 MHz) reveals red MB to be identical to
MVB and not MB, nor any hydroxy adduct. Thus, the NMR spectra of
MB, red MB and MVB depicted in Fig. 2 clearly indicate the
conversion of a symmetrical molecule (MB, lower NMR spectrum)
to an unsymmetrical molecule possessing two unique and clearly
distinguishable AMX spin systems entirely consistent with MVB
and in contrast to what would be predicted for the symmetrical
MB-OH adduct 2. Consistent with the above is the mass spectrum of
the red MB in toluene solution which revealed the predominant
presence of MVB (molecular ion peak at m/z 256) with some azure B
(5; m/z peak at 270).
3.3. The nature of heterocyclic nitrogen in MB and analogous dyes
Based on our foregoing investigations with MB and the lack of
information of any product with a structure MB-OH (2), we
remained interested in the question of the electronegativity of the
heterocyclic nitrogen in MB. Fig. 3 shows the main valence bond
resonance forms contributing to the structure of MB. A central
contention [7,12] underlying the claimed formation of MB-OH (2)
was that the central nitrogen in MB is electron deficient. The
implication is that the mesomer N, characterized by a positive
charge localized on the heterocyclic N connecting two fully
benzenoid rings, is a significant contributor to the overall structure.
However, density functional calculations and application of both
the NBO and electrostatic potential (ESP) methods reveal this not to
be the case. Rather, it appears the central nitrogen is electron rich
relative to the remainder of the molecule. Thus, while the magni-
tude of the charges calculated by each method varies, both methods
reveal that the S atom carries a neutral (ESP: þ0.04e) to partial
positive (NBO: þ0.52e) charge, reflecting the importance of
mesomer S. Importantly, with both methods the heterocyclic N
atom of MB has an overall partial negative charge (NBO: ꢁ0.38e,
ESP: ꢁ0.79e) as illustrated by the calculated partial charges from
the NBO method in Fig. 4. In valence bond terms mesomer N is
a minor contributor, while mesomers S and D1/D2 more closely
resemble the structure of MB.
Overall, these experimental findings show that the red MB
species introduced by Plater [12] and subsequently claimed by Pal
and co-workers [7] to be the N-hydroxy species MB-OH (2) is in fact
MVB (4), possibly containing some azure B (5), generated by adding
a high concentration of base (0.1 M NaOH) to an aqueous solution of
MB, through a simple, well-established [30] hydrolysis reaction.
In the meantime, Plater has indicated that the red material to
which he assigned the MB-OH structure 2 [12] is in fact Azure B
One of our early reasons for questioning the claim for a N-
hydroxy adduct structure 2 for MB-OH was experience of simple
colourestructure relationships. A molecule with the structure 2 of
Fig. 2. 1H NMR spectra of (top to bottom) purchased MVB (Aldrich), red MB (both recorded in CCl4 using Bruker Avance-III 600 NMR spectrometer) and methylene blue (Acros)
(recorded in D2O using Bruker Avance/DRX 500 NMR spectrometer).

A. Mills et al. / Dyes and Pigments 88 (2011) 149e155
153
heterocyclic N atom (pKBHþ ¼ ꢁ0.26) [40,41]. Alternative protonation
of a pendant Me2N group would lead to a hypsochromic shift, and is
discounted.
N
S
N
S
+
+
Me2N
NMe2
NMe2
Me2N
All the above evidence suggests that although a valid contrib-
utory valence bond MB structure can be written with electron-
deficient nitrogen, N in Fig. 3, this does not represent the actual
structural situation of MB. In fact, the heterocyclic nitrogen atom in
N is now a formal 6-electron centre, analogous to a diarylnitrenium
ion (a known, highly reactive species [42]). Electron deficiency due
to a sub-octet electron configuration at a relatively electronegative
element such as nitrogen is energetically disfavoured, and so even if
mesomer N appears acceptable on paper, it is in fact unrepresen-
tative of MB. The same conclusion appears to hold across other dye
classes which include formally 6-electron positive sp2 nitrogen
centres. As far as we are aware, there are no reports of N-hydroxy
derivatives of analogues such as diazines, quinone imines (typified
by Bindschedler’s Green (C.I. 76125; 9), which hydrolyses via
successive attack at the two terminal C-NMe2 groups [43]), aza-
cyanines, and so on. Furthermore, simple MO calculations of such
D 2
D 1
..
+
N
N
NMe2
Me2N
Me2N
S
+
S
NMe2
S
N
Fig. 3. Major valence bond resonance structures of MB. Alternative Kekule structures
of benzenoid rings and charged-carbon mesomers are not shown.
species always indicate
a relatively electronegative nitrogen,
MB-OH would not be expected to be red: the delocalized conju-
gation present in MB is removed in MB-OH, just as it is in
(colourless) leuco-MB (3), and MB-OH is equally unlikely to be
coloured. In much earlier classical dyeestructure studies, Lewis
et al. speculated on the formation of a product with structure
MB-OH (2) by the addition of water to MB as an intermediate in
their chemistry, and they also noted it would be colourless [37].
Subsequently we have calculated the absorption maximum for
a species with the MB-OH structure 2. After initial geometry opti-
mization, TD-DFT calculations reveal it would have an absorption
maximum in the near UV at 351 nm, confirming its essential lack of
colour. Similar calculations of the visible spectra of other known
thiazine dyes reproduce their experimental absorption maxima
with reasonable reliability (Table 1), demonstrating the suitability
of the level of theory used for this work [38].
Our own wide-ranging literature search for precedent of
hydroxide anion addition at imine-type nitrogen has turned up no
example [39]. Whilst Plater [12] only speculated on the electroneg-
ative nature of the heterocyclic nitrogen of MB and the subsequent
formation of MB-OH, albeit without any direct structural character-
ization, it is apparent that this speculation has been over-interpreted
by others [7,11] and has given credence to electron-deficient
heterocyclic nitrogen in MB when none is due.
bearing a partial negative charge [44]. In contrast, when the
nitrogen centre of interest is replaced by positively charged sp2
carbon, hydroxide addition at that centre is well known. Examples
include diaryl carbenium ions (e.g. Michler’s hydrol, 10), their tri-
aryl carbenium ion analogues (e.g. Malachite Green (C.I. Basic
Green 4; 11), and Crystal Violet (C.I. Basic Violet 3; 12)) and the
cyclised acridine, xanthene and thioxanthene series [45]. Unlike
electronegative nitrogen, the more electropositive carbon is more
amenable to supporting a positive charge in an electron-deficient
6-electron configuration, and thus can be a site for attack by
negative nucleophiles.
X
+
N
+
Me2N
NMe2
Me2N
(10) X = H
NMe2
(9)
(11) X = Ph
(12) X = 4-Me2NC6H4
3.4. Other thiazines
If there is no experimentalevidence forelectron-deficient nitrogen
in MB, is there experimental support for this centre being electron
rich, as calculated? Protonation of MB in strong acid is well known to
give a bathochromic shift in lmax, from 664 nm for MB in water to
746 nm in aqueous acidic MB, consistent with protonation at the
The purities of the other major thiazines, namely Azures A, B, C
and thionine (5e8; see Table 1), were assessed using LC-MS and the
results are summarized in Table 2. Thus, whereas the commercial
sample of thionine (8) was very pure (99þ%) and azure B (5) of
a reasonable purity (ca. 86%), the other two thiazines were gross
mixtures. Most notably, the sample of azure A (6) was of very low
purity (ca. 21%) and far from the supplier’s claimed value of 96%.
However, for the general purpose of demonstrating the formation of
coloured lipophilic species upon treatment with alkali the dyes
Table 2
Composition of commercial samples of thiazine dyes used in this study, determined
by hplc.
Commercial
thiazine
%MB (1)
%AB (5)
%AA (6)
%AC (7)
%TH (8)
MB (1)
AB (5)
AA (6)
AC (7)
TH (8)
MVB (4)
91
3
36
0
8.7
86
28
7
<1
10
27
47
0
<1
1
8
38
0
<1
0
1
7.5
99þ
0
0
Fig. 4. NBO charge distribution on the PCM/B3LYP electronic ground state of MB. See
Computational Methods for details.
84% MVB, no other thiazines present

154
A. Mills et al. / Dyes and Pigments 88 (2011) 149e155
were used as received. Thus, these four thiazine dyes were treated
with an aqueous pH 13 solution, and photographs of the solutions (i)
before, (ii) directly after (with shaking), and (iii) 5 h after mixing
with an identical volume of toluene, are shown in Fig. 5. Given the
reports [7,12] of a coloured, lipophilic hydroxy adduct, some might
be tempted to assume that all the thiazine dyes form such species
and assign the highly coloured toluene solutions arising from
alkaline treatment illustrated in Fig. 5 as being due to the N-hydroxy
adducts of the associated thiazine. However, this work now shows
that the reported red MB is, in fact, MVB and that the N-hydroxy
adduct of any thiazine should be largely colourless. Thus, the
alternative, less glamorous, but well-established explanation for the
results illustrated in Fig. 5 still stands, i.e. that the other thiazines are
simply deprotonated by alkali to form highly coloured lipophilic
OH
N
N
S
+
NMe2
Me2N
Me2N
S
NMe2
(1) MB
(2) MB-OH
aq NaOH
N
S
N
S
+
O
NMe
Me2N
Me2N
Me2N
Me2N
HO-
H+
HO-
H+
N
S
N
S
+
N
H
+
OH
Me
(5) AB (minor)
(4) MVB (major)
Fig. 6. Summary of reactions of MB with alkali.
species. The results in Fig. 5 also reveal that the thionine (8) and
azure C (7) pH 13 aqueous solutions (before mixing with toluene) are
red, whereas those for azure A (6) and azure B (5) are purple. This
feature arises because the pKa for the thiazine dyes (see Table 1)
decreases with decreasing degree of methylation, so that, for
example, at pH 13 most (99%) of the thionine will be in its red,
lipophilic deprotonated, free base form, whereas for Azure B at least
11.2% will still be in its purpleeblue protonated form [31].
The calculated lmax values in toluene of the deprotonated forms
of the thiazine dyes, along with those found experimentally, are
listed in Table 1. The experimental and calculated values compare
very well (especially given the impure natures of azure A and azure
C) and strongly support the deprotonation mechanism. Others [46]
have used NMR spectroscopy to show that the yellow/orange
lipophilic form of thionine, produced by treatment with alkali, is
the deprotonated form of the parent dye.
4. Conclusions
The chemistry discussed in this paper is summarized in Fig. 6.
Basic hydrolysis of MB leads mainly to the production of MVB, 4,
and the visible spectral properties of MVB in six solvents and on
a tlc plate match very closely the reported properties of what has
been claimed [7] to be an N-hydroxy adduct of MB, MB-OH (2).
However, the calculated visible spectrum of MB-OH is reliably
predicted to be colourless, not red, and the incorrectly assigned site
of hydroxide attack in MB, the heterocyclic N atom, is not electron
deficient as suggested as a reason for MB-OH formation. We have
been unable to find any literature precedent for hydroxide attack at
any sp2-hybridised nitrogen atom, and all the experimental
evidence points to MVB (4) as the main product of alkaline
hydrolysis of MB, and not MB-OH (2).
It is always difficult to prove a negative. We are not claiming that
the nitrogen heterocyclic site in MB or any similar nitrogen can
never be electrophilic in appropriate circumstances. But as far as we
can determine this has not yet been demonstrated, and available
evidence such as the results of calculation and site of protonation
indicates the opposite.
Fig. 5. Photographs of a range of freshly made 10ꢁ4 M thiazine dye solutions in 0.1 M
NaOH (from left to right): before, directly after, and 5 h after mixing and shaking with
an equal volume of toluene.

A. Mills et al. / Dyes and Pigments 88 (2011) 149e155
155
[10] Mills A, Hazafy D, Parkinson JA, Tuttle T, Hutchings MG. J Phys Chem A
2009;113:9575e6.
[11] Pal T. J Phys Chem A 2009;113:9577.
[12] Plater MJ. ARKIVOC; 2003:37e42.
[13] Marshall PN, Lewis SM. Stain Technol 1976;50:375.
[14] Barone V, Cossi M. J Phys Chem A 1998;102:1995.
[15] Cossi M, Rega N, Scalmani G, Barone V. J Comput Chem 2003;24:669.
[16] Becke AD. Phys Rev A 1988;38:3098.
[17] Becke AD. J Chem Phys 1993;98:5648.
[18] Hertwig RH, Koch W. Chem Phys Lett 1997;268:345.
[19] Lee CT, Yang WT, Parr RG. Phys Rev B 1988;37:785.
[20] Stephens PJ, Devlin FJ, Chabalowski CF, Frisch MJ. J Phys Chem 1994;98:11623.
[21] Vosko SH, Wilk L, Nusair M. Can J Phys 1980;58:1200.
[22] Hariharan PC, Pople JA. Theor Chim Acta 1973;28:213.
[23] Krishnan R, Binkley JS, Seeger R, Pople JA. J Chem Phys 1980;72:650.
[24] Bauernschmitt R, Ahlrichs R. Chem Phys Lett 1996;256:454.
[25] Casida ME, Jamorski C, Casida KC, Salahub DR. J Chem Phys 1998;108:4439.
[26] Stratmann RE, Scuseria GE, Frisch MJ. J Chem Phys 1998;109:8218.
[27] Frisch MJ, Trucks GW, Schlegel HB, Scuseria GE, Robb MA, Cheeseman JR, et al.
Gaussian 03, revision E.04. Wallingford CT: Gaussian, Inc.; 2004.
[28] Reed A, Weinstock RB, Weinhold F. J Chem Phys 1985;83:735.
[29] Abbott DC. Analyst 1962;87:286.
The other major cationic thiazine dyes, which contain one or
more amine-attached hydrogen atoms, also do not form hydroxy
adducts, but instead are deprotonated in 0.1 M NaOH solution to
very differently coloured (lmax hypsochromic shift in aqueous
solution typically ¼ 100e150 nm), red or orange lipophilic, free base
forms of the original dye (shown for azure B, 5, in Fig. 6) It is also
important to note that similar product formation and spectral
changes would be expected for oxazine and phenothiazine dyes.
Thus the claim [7] that Nile Blue (C.I. Basic Blue 12) forms a coloured,
lipophilic N-hydroxy adduct, rather than a deprotonated lipophilic
species (like its near thiazine equivalent, Azure A, 6), appears as
unlikely as that of a coloured, N-hydroxy adduct of MB.
Acknowledgement
We thank Dr M.J. Plater of Aberdeen University for helpful
discussions about his own studies on methylene blue and its
interaction with hydroxide, especially clarification of the red
material he obtained from MB at pH 11.
[30] Holmes WC, Snyder EF. Stain Technol 1929;4:7.
[31] Bonneau R, Faure J, Joussot-Dubien J. Talanta 1967;14:121.
[32] The discoverer’s name is frequently appended when naming the thiazine-
based Methylene Violet, to distinguish it from other species with the same or
similar name e.g. Methylene Violet 3RAX, which has an unrelated safranine
structure, 3-amino-7-(diethylamino)-5-phenyl phenazinium chloride; see
Aldrich Catalogue 307505.
References
[33] Otsuki S, Taguchi T. Bull Chem Soc Jpn 1996;69:2525.
[34] Plater MJ. Private communication, February 2010.
[1] Travis AS. The rainbow makers. London: Associated University Presses; 1993.
[2] The merck index. Rahway: Merck & Co., Inc.; 1989.
[35] Mills A, Hazafy D. unpublished results.
[3] http://en.wikipedia.org/wiki/Methylene_blue accessed 01 2010.
[4] Murray AD, Stuff RT, Ahearn TS, Bentham P, Seng KM, Wischik C. Alzheimer’s
Demen 2008;4:T786.
[5] Wischik CM, Rickard JE, Harrington CR, Horsley D, Storey JMD, Marshall C,
Sinclair JP, Tan HW. Methods of chemical synthesis and purification of dia-
minophenothiazinium compounds including methylthioninium chloride, U.S.
patent 2006/0287523 A1 published Dec 21st 2006.
[36] Marshall PN, Lewis SM. Biotech Histochem 1975;50:375e81.
[37] Lewis GN, Goldschmid O, Magel TT, Bigeleisen J. J Am Chem Soc 1953;65:1150.
[38] A recent paper examines the use TD DFT calculations of excitation energies of
dyes, including MB. Fabian J Dyes Pigm 2010;84:36e53.
[39] Beilstein database v.2008/04, via CrossFire Commander v.7.1,
Information Systems.
Ó
Elsevier
[40] Nishikiori H, Nagaya S, Tanaka N, Katsuki A, Fujii T. Bull Chem Soc Jpn
1999;72:915e21.
[41] Lewis GN, Bigeleisen J. J Am Chem Soc 1943;65:1144e50.
[42] Falvey DE. J Phys Org Chem 1999;12:589e96.
[43] Griffiths J, Cox R. Dyes Pigm 1999;42:29e34; Möhlau R. Ber Deutschen Chem
Ges 1884;18:2915.
[6] http://www.examiner.com/x-11432-Pittsburgh-Medical-Technology-Examinerw
y2010m1d9-Mehtylene-blue-may-hold-the-cure-for-neurodegenerative-diseases-
such-as-Alzheimers-disease?cid¼edition-rss-Pittsburgh; January 09 2010.
[7] Basu S, Panigrahi S, Praharaj S, Ghosh SK, Pande S, Jana S, et al. J Phys Chem A
2007;111:578.
[8] Thien TV [chapter 3]. In: Muthyala R, editor. Chemistry and applications of
leuco dyes. New York: Plenum Press; 1997. p. 67.
[44] Unpublished results.
[45] Zollinger H. Color chemistry [chapter 4]. 3rd ed. Zürich: Vhca and Wiley-VCH;
2003.
[46] Nicotra VE, Mora MF, Iglesias RA, Baruzzi AM. Dyes Pigm 2008;76:315.
[9] Adamcikova L, Pavlikova K, Sevcik P. React Kinet Catal Lett 2000;69:91;
Holmes WC, French RW. Stain Technol 1926;1:17; MacNeal WJ, Killian JA.
J Am Chem Soc 1926;48:740.