Molecular implementation of sequential and reversible logic through photochromic energy transfer switching

Article abstract of DOI:10.1002/chem.201100027

Photochromic spiropyrans modified with fluorophores were investigated as molecular platforms for the achievement of fluorescence switching through modulation of energy transfer. The dyads were designed in such a way that energy transfer is only observed for the open forms of the photochrome (merocyanine and protonated merocyanine), whereas the closed spiropyran is inactive as an energy acceptor. This was made possible through a deliberate choice of fluorophores (4-amino-1,8-naphthalimide, dansyl, and perylene) that produce zero spectral overlap with the spiro form and considerable overlap for the merocyanine forms. From the Foerster theory, energy transfer is predicted to be highly efficient and in some cases of 100 % efficiency. The combined switching by photonic (light of I>530 nm) and chemical (base) inputs enabled the creation of a sequential logic device, which is the basic element of a keypad lock. Furthermore, in combination with an anthracene-based acidochromic fluorescence switch, a reversible logic device was designed. This enables the unambiguous coding of different input combinations through multicolour fluorescence signalling. All devices can be conveniently reset to their initial states and repeatedly cycled. Copyright

Full text of DOI:10.1002/chem.201100027

DOI: 10.1002/chem.201100027
Molecular Implementation of Sequential and Reversible Logic Through
Photochromic Energy Transfer Switching
Patricia Remꢀn,[a] Martin Hammarson,[b] Shiming Li,[b] Axel Kahnt,[b]
Uwe Pischel,*[a] and Joakim Andrꢁasson*[b]
In memoriam Rafael Suau
Abstract: Photochromic spiropyrans
modified with fluorophores were inves-
tigated as molecular platforms for the
achievement of fluorescence switching
through modulation of energy transfer.
The dyads were designed in such a way
that energy transfer is only observed
for the open forms of the photochrome
(merocyanine and protonated merocya-
nine), whereas the closed spiropyran is
inactive as an energy acceptor. This
was made possible through a deliberate
choice of fluorophores (4-amino-1,8-
naphthalimide, dansyl, and perylene)
that produce zero spectral overlap with
the spiro form and considerable over-
lap for the merocyanine forms. From
the Fçrster theory, energy transfer is
predicted to be highly efficient and in
some cases of 100% efficiency. The
combined switching by photonic (light
of l>530 nm) and chemical (base)
inputs enabled the creation of a se-
quential logic device, which is the basic
element of a keypad lock. Further-
more, in combination with an anthra-
cene-based acidochromic fluorescence
switch, a reversible logic device was de-
signed. This enables the unambiguous
coding of different input combinations
through multicolour fluorescence sig-
nalling. All devices can be conveniently
reset to their initial states and repeat-
edly cycled.
Keywords:
logic gates
energy transfer
· photochromism
·
·
spiro compounds · switches
Introduction
as conveniently readable outputs.[1–3] Photophysically, these
systems are often based on well-understood energy-transfer
and electron-transfer processes.[13] This facilitates the ration-
al design of such switches and enables effective control of
the fluorescence output. The same excited state pathways
were found to be modulated by the interaction of fluoro-
phores with photochromic molecules in appropriately de-
signed dyads, leading to novel photonic switches.[14,15] Photo-
chromes themselves can be understood as bistable or some-
times multistable systems, in which the states involved show
considerable differences in properties such as UV/Vis ab-
sorption and redox potentials. One of the commonest appli-
cations for information processing with photochromes is
their exploitation as molecular memories.[16–23] Notably, even
more complex logic operations such as those of multiplex-
ers,[24] demultiplexers,[25] encoders, decoders,[26] half-
adders[27,28] and keypad locks[29] have been devised through
combining the action of several photochromes. This has
opened up a strategy for the design of all-photonic molecu-
lar logic devices using exclusively optical input and output
signals.
The idea of using molecular systems for mimicking binary
information processing has attracted considerable interest in
recent years.[1–4] On the one hand, such molecular entities
could be imagined as building blocks for molecular comput-
ers. In practice, however, problems such as the concatena-
tion of several logic devices create bottlenecks for such ap-
plications.[4] On the other hand, the utility of molecular logic
is not restricted to computing,[5] but has, for example, emer-
gent interest for smart materials,[6] pro-drug activation,[7,8] la-
belling of micro-objects,[9] multiparameter sensing[10] and
molecular diagnostic tools.[11,12]
Many of the numerous molecular logic devices reported
so far use chemical signals as inputs and fluorescence signals
[a] P. Remꢀn, Dr. U. Pischel
Center for Research in Sustainable Chemistry (CIQSO) and
Department of Chemical Engineering, Physical Chemistry,
and Organic Chemistry
University of Huelva, Campus de El Carmen, s/n
21071 Huelva (Spain)
Fax : (+34)959-21-99-83
One of the most popular classes of photochromic switches
are spiropyrans such as 1-SP (Scheme 1). For such a com-
pound three different states are observed: SP (spiro form),
ME (merocyanine form) and MEH+ (protonated merocya-
nine form). Thanks to the pronounced differences in the ab-
sorption spectra of these states, energy transfer switching
has been used to control the fluorescence intensities of at-
tached fluorescent units in spiropyran-based supramolecular
E-mail: uwe.pischel@diq.uhu.es
[b] M. Hammarson, Dr. S. Li, Dr. A. Kahnt, Prof. Dr. J. Andrꢁasson
Department of Chemical and Biological Engineering
Physical Chemistry, Chalmers University of Technology
41296 Gçteborg (Sweden)
Fax : (+46)31-772-38-58
E-mail: a-son@chalmers.se
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201100027.
6492
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 6492 – 6500

FULL PAPER
SP (yields of 54, 57 and 68%, respectively). The fluorophore
amine derivatives were prepared as described in the Sup-
porting Information. The fluorophore model compounds (2-
M, 3-M, and 4-M) were obtained by direct acetylation of the
amines with acetyl chloride or acetic anhydride
(Scheme 2b).
Photochromic behaviour of the model spiropyran 1-SP:
Compound 1-SP can be considered a model spiropyran in
our study, because it resembles the photochrome part of all
the dyads investigated here. The corresponding absorption
spectra of all three forms of 1 are displayed in Figure 1. As
shown in Scheme 1, irradiation of the SP form (lmax
=
Scheme 1. Structures of 1-SP, 1-MEH+ and 1-ME and their interconver-
sion through photonic and/or chemical inputs.
constructs (dyads etc.). This has been accomplished in liquid
solution,[30] in polymers,[31] on nanoparticles[32] and in live
cells.[33] As shown in Scheme 1, spiropyran photochromes
combine the use of photonic and chemical input signals.
In this work, the covalent linking of a nitrospiropyran
with carefully chosen fluorophores (dansyl, 4-amino-1,8-
naphthalimide and perylene) has enabled fluorescence
switching through the modulation of energy transfer pro-
cesses through the action of chemical and photonic inputs.
The rational design and comprehensive interpretation of the
switching behaviour has been used here to achieve a fluores-
cent sequential logic device, which is the basic function of a
molecular keypad lock.[29,34–38] In combination with an an-
thracene fluorescence switch, a logically reversible gate was
also devised.[39,40]
Figure 1. Absorption spectra of 1-SP (dotted line), 1-MEH+ (dashed
line) and 1-ME (solid line) in acetonitrile.
341 nm, e=7500mꢀ1 cmꢀ1) with UV light (l=302 nm, 210 s)
leads to ring opening and the generation of the ME form (1-
ME). This is characterized by a substantially red-shifted ab-
sorption band at lmax =565 nm (e=36800mꢀ1 cmꢀ1). The dis-
tribution between the two forms in the photostationary state
(PSS) was determined as 39% SP and 61% ME. The 1-ME
form reverts to 1-SP either thermally (t=6.1 min at 208C)
or, more rapidly, on visible light irradiation. The time re-
quired for the photoinduced isomerization depends, of
course, on the light intensity. Here, irradiation at l>530 nm
with a light power density of about 18 mWcmꢀ2 required
210 s exposure time for the ME to SP conversion.
The addition of trifluoroacetic acid (TFA, 1 equiv) to 1-
ME generated the protonated MEH+ form (1-MEH+),
which was accompanied by a blue shift of the long-wave-
length absorption maximum to 404 nm (e=22700mꢀ1 cmꢀ1).
This process can be reversed by addition of a strong phos-
phazene base (P2-Et, 1 equiv). The open forms can be quan-
titatively transformed back to the closed SP form by irradia-
tion with light of l>420 nm (ME and MEH+) or, for the se-
lective conversion of ME, of l>530 nm.
Results and Discussion
Synthesis of the spiropyran–spacer–fluorophore dyads (2-SP,
3-SP, 4-SP) and model compounds: The complete synthesis
sequence for the dyads is shown in Scheme 2a. The prepara-
tion of 1-SP was accomplished by starting with the commer-
cially available 2,3,3-trimethylindolenine and following a
three-step procedure, which was slightly modified with re-
spect to the literature (see the Supporting Information).[41,42]
Alkylation of the indolenine with methyl 4-bromobutyrate
gave the corresponding indolinium salt. In the second step
this salt was treated with 5-nitrosalicylaldehyde to yield the
nitrospiropyran in its methyl ester form. The ester was final-
ly treated with sodium hydroxide in tetrahydrofuran, fol-
lowed by acidification with citric acid to yield 1-SP. This
compound was then further functionalized with the corre-
sponding fluorophores through formation of amides. The
best yields were obtained by prior acid activation by the
NHS ester method (see the Supporting Information). Treat-
ment of the NHS ester with the corresponding fluorophore
amine derivatives yielded the target dyads 2-SP, 3-SP and 4-
Photophysical properties and photochromic switching of the
dyads: All relevant spectral data for the dyads in their dif-
Chem. Eur. J. 2011, 17, 6492 – 6500
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6493

U. Pischel, J. Andrꢁasson et al.
Scheme 2. a) Synthesis of 1-SP and its functionalization with fluorophores (2-SP, 3-SP, 4-SP). DCC=N,N’-dicyclohexylcarbodiimide. b) Preparation of
fluorophore model compounds (2-M, 3-M, 4-M).
ferent forms (SP, ME, MEH+) and for the fluorophore
models are compiled in Table 1. Figure 2 compares the ab-
sorption spectra of the dyad 2-SP and the corresponding
model compounds 1-SP and 2-M. The combination of the
spectra of the models yields the spectrum of the dyad, which
rules out significant ground-state interactions. The same ob-
servations were made for the other two dyads (see the Sup-
porting Information). Notably, the fluorophore emission of
the dyads in their SP forms is considerably quenched (ca.
90%) relative to the emission of the model chromophores
(Figure 2 for 2-M versus 2-SP). The underlying fluorescence
quenching process is dynamic in nature, as has been verified
by the observation of the same effect in fluorescence life-
time measurements. For the 4-amino-1,8-naphthalimide fluo-
rophore, for example, the quantum yield (Ff) drops from
0.44 for 2-M to 0.056 for 2-SP and the lifetime (tf) is 10.6 ns
for 2-M and 1.40 ns for 2-SP (see Table 1). The mechanism
for this fluorescence quenching is assumed to be photoin-
duced electron transfer (PET). Electronic energy transfer
(EET) can be excluded because the dyads were designed to
give a zero spectral overlap integral between the fluoro-
phore emission and the SP absorption spectrum (see below).
To validate the hypothesis of PET, the driving force (DGPET
)
for the oxidative and reductive paths was calculated by ap-
plication of the Rehm–Weller equation [Eq. (1)].[30,43–46]
*
DGPET ¼ Eox ꢀ Ered ꢀ E þ C
ð1Þ
In all cases, PET with oxidation of the fluorophore is exer-
gonic and therefore thermodynamically feasible: DGPET,ox
=
ꢀ0.31, ꢀ0.91 and ꢀ0.76 eV for 2-SP, 3-SP and 4-SP, respec-
tively (errorꢁ0.1 eV). Reductive PET was also considered,
but with the exception of 2-SP (DGPET,red =ꢀ0.15ꢁ0.1 eV)
this path is endergonic (DGPET,red >0). Notably, despite the
6494
www.chemeurj.org
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 6492 – 6500

Molecular Implementation of Sequential and Reversible Logic
FULL PAPER
Table 1. Photophysical properties of model compounds and dyads.
As is discussed below, the
dyads described here were de-
signed to yield fluorescence
switching dependent on the
state of the photochrome
moiety. This idea is based on
the control of very efficient
EET (see below) being sup-
pressed in the SP forms of the
dyads and activated for the
other two forms (ME and
MEH+): that is, ON/OFF fluo-
rescence switching upon photo-
induced ring opening of the SP
forms. It should be stressed that
to achieve the maximum ON/
OFF discrimination, EET in the
ME and MEH+ forms of the
dyads should be much more ef-
ficient than PET in the corre-
sponding SP forms; from the
lifetimes in Table 1 it can clear-
ly be seen that this condition is
fulfilled.
[c]
[g]
labs,max [nm]
lf,max
Ff[a]
tf[b]
FPET
PSS[d]
Q [%][e]
R0
FEET
(e/mꢀ1 cmꢀ1
)
[nm]
[ꢃ][f]
2-M
2-SP
2-ME
2-MEH+
3-M
416 (10400)
512
0.44
10.6 ns
1.40 ns
13 ps
15 ps
12.6 ns
925 ps
8 ps
416 (11000)
512
0.056
0.035[j]
0.035[j]
0.25
0.015
0.008[j]
0.008[j]
0.76
0.87
55
45
0
47
20
[h]
[i]
38
38
1.0
0.85
[h]
[i]
338 (4600)
518
3-SP
341 (11000)
518
0.94
0.90
54
46
0
43
19
[h]
[i]
3-ME
3-MEH+
4-M
44
44
1.0
0.80
[h]
[i]
9 ps
439 (33600)
446
4.0 ns
560 ps
16 ps
16 ps
4-SP
439 (32100)
447
0.079
0.034[j]
0.034[j]
58
42
0
38
35
[h]
[i]
4-ME
57
57
1.0
0.99
4-MEH+
[h]
[i]
[a] Fluorescence quantum yields of the model compounds measured with quinine sulfate in H2SO4 (0.05m) as
standard (Ff =0.55), 20% error. The values for the SP dyads were determined relative to those of the models.
[b] Fluorescence lifetimes, 10% error. [c] Quantum yields for PET, calculated by Equation (2). [d] Distribution
of spiropyran and merocyanine isomers (%SP/%ME) in the PSS (l=302 nm). [e] Fluorescence quenching ob-
served for the PSS (SP!ME/MEH+). [f] Critical radius of electronic energy transfer (Fçrster mechanism),
calculated by Equation (3). [g] Quantum yields for electronic energy transfer estimated from the Fçrster for-
mulation, with the assumption of a distance of 15 ꢃ between fluorophore and photochrome. [h] The long-
wavelength absorption maxima of the protonated merocyanine (MEH+) and the merocyanine (ME) forms of
the dyad are determined by the photochrome part (see text for data for the model compounds 1-MEH+ and
1-ME). [i] The fluorescence maximum remains unaltered with respect to the fluorescence observed for the SP
form of the corresponding dyad. [j] Apparent fluorescence quantum yield, determined from the fluorescence
quenching in the PSS (irradiation of the SP form with 302 nm light) and subsequent addition of TFA (1 equiv,
for the MEH+ form).
The fluorophores play the
role of energy donors, whereas
the photochrome moiety is the energy acceptor. According
to the Fçrster theory of energy transfer, the efficiency
(FEET) of this process depends crucially on the spectral over-
lap integral (J) of the donor emission and the acceptor ab-
sorption [Eqs. (3) and (4)]. With Equation (3) the critical
distance for energy transfer (R0) can be calculated (the ori-
entation factor k2 was assumed to be 2/3). This distance cor-
responds to a 50% EET efficiency. For all dyads, considera-
ble R0 values of 38–47 ꢃ were found for their ME forms,
whereas somewhat smaller values were determined for their
MEH+ forms (R0 =19–35 ꢃ). Importantly, and as intended,
R0 =0 ꢃ applies for the SP forms of the dyads, because the
spectral overlap integral is zero. This means that EET does
not occur in this photochromic state of the dyads. With the
real distance between donor and acceptor the values for
FEET can be calculated [Eq. (4)]. Because of the conforma-
tional flexibility of the linker, however, the actual distance
between donor and acceptor is expected to follow a distribu-
tion function. For this reason we assume an estimated value
of the maximum distance, corresponding to the most extend-
ed conformation, with R about 15 ꢃ. Most probably the real
average distances are even smaller, as can be judged from
the occurrence of PET in the SP forms of the dyads (see
above). This process would not be efficient (see Table 1) for
a distance of 15 ꢃ. In any case, this maximum distance is
still significantly smaller than R0 for both of the open forms.
The lower limit values for EET efficiency thus show that for
the ME form this process is expected to be quantitative
(FEET =1) for all fluorophores, whereas for the protonated
merocyanine form (MEH+) considerably high values of
FEET ꢃ0.8 should still result.
Figure 2. Absorption spectra (three spectra on the left) of 2-M (solid
line), 1-SP (dashed line) and 2-SP (dotted line) in acetonitrile. On the
right (above 450 nm) the fluorescence spectra of optically matched solu-
tions of 2-M (solid line) and 2-SP (dotted line) in acetonitrile are shown
(lexc =417 nm).
exergonic thermodynamics in the case of 2-SP, the occur-
rence of reductive electron transfer is unlikely, because elec-
tron donors linked to the “imide side” of the chromophore
do not quench the fluorescence of aminonaphthali-
mides.[47,48] Oxidative PET is therefore a very likely possibil-
ity for the observed fluorescence quenching in the SP dyads.
With the assumption that PET is the only competitive fluo-
rescence quenching process in the SP dyads, the quantum
yield can be given with FPET ꢂ0.9 [Eq. (2)].
Ff;SP
FPET ¼ 1 ꢀ
ð2Þ
Ff;M
Chem. Eur. J. 2011, 17, 6492 – 6500
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6495

U. Pischel, J. Andrꢁasson et al.
1 FDðnÞeAðnÞdn
Z
9 ln 10k2F
The other two dyads showed the same lifetime quenching
trend (Table 1). In other words, each dyad molecule, on con-
version from the SP to the open ME/MEH+ forms, shows
quantitative fluorescence quenching.
With the complete characterization of the photochrome-
modulated fluorescence switching available, we set out to
achieve the following advanced logic functions: sequential
molecular logic[49,50] and reversible logic.[39,40]
R60 ¼
D J with J ¼
ð3Þ
ð4Þ
128p5n4NA
n4
0
1
FEET
¼
ꢀ
ꢁ
6
R
1 þ
R0
The expected ON/OFF fluorescence switching upon con-
version of the SP form to the ME form was experimentally
verified for the irradiation of the dyads at l=302 nm. The
conversion 2-SP!2-ME, for example, resulted in fluores-
cence quenching of 38%. This effect is even somewhat
more pronounced for the dansyl- and perylene-substituted
spiropyrans (3-SP and 4-SP), for which the emission is re-
duced by about 44 and 57%, respectively, upon merocya-
nine formation. Clearly, the dynamic range of the fluores-
cence switching is dictated by the PSS distribution, which
was found to be around 55:45 (%SP/%ME) for all three
dyads. The similarity of the quenching and the ME contribu-
tion to the PSS is consistent with the prediction of very effi-
cient energy transfer (see above). Comparison of the PSS
distribution for the dyads and the spiropyran model com-
pound 1-SP (%SP/%ME 39:61) reveals that the photochro-
mic ring opening process is somewhat affected in the dyads,
which is possibly the result of photochrome excited state
quenching by the fluorophores (e.g., through PET). Quanti-
tative fluorescence quenching by EET in the dyads in their
open forms (ME and MEH+) was also verified by fluores-
cence lifetime measurements (see Table 1). For the example
of the dansyl-substituted photochrome (3-SP) a drastic re-
duction of the lifetime from 925 ps for the SP form to a few
picoseconds (8–9 ps for the component of shortest lifetime)
for the ME and MEH+ forms was observed (see Figure 3).
Implementation of sequential logic through chemical and
photonic inputs: The outcome of a sequential logic opera-
tion is dependent not only on the correct combination of the
inputs, but also on the order of their application. This im-
plies the existence of a memory effect. Prominent examples
of logic switching of this type are keypad locks and SR
latches,[51] for which a restricted number of molecular exam-
ples are known. Keypad locks are devices that activate only
if the right input combination is applied in a certain se-
quence. With one exception,[29] the molecular keypad locks
reported to date rely on all-chemical input signals.[35–38] This
could imply problems for the repeated use of the molecular
system, which would require mechanisms for resetting and
recycling.
As shown in Scheme 1, not only do spiropyran photo-
chromes offer the possibility of light-induced switching, but
their open forms can also be interconverted by protonation/
deprotonation. Hence, three spectrally differentiated states
(SP, ME, MEH+) can be observed for these photochromes.
Effectively, a chemical (sensing) input and a remote-ad-
dressing photonic input are combined in a “hybrid ap-
proach” (neither all-chemical nor all-photonic). This can be
used to devise comprehensive logic operations,[52–57] beyond
the bistable photoswitching of spiropyrans (SP!ME and
vice versa). It should be noted here that other classical pho-
tochromes, such as dithienylethenes or fulgimides, have only
two differentiated states (closed and open form). Thus, to
achieve complex logic functionality with these photo-
chromes, several distinct units are commonly combined in a
supermolecule conjugate.[24–26,28,29]
For the spiropyran–fluorophore dyads investigated here,
fluorescence constitutes a convenient output signal (O1). As
the initial state of our sequential logic device, the MEH+
form of the dyad was chosen. One input (In2) is light of l>
530 nm, at which the initial MEH+ state shows no absorp-
tion, but the ME form does absorb (see Figure 1). To gener-
ate the SP form, the MEH+ form has to be converted into
the ME form before irradiation with light of l>530 nm (see
Scheme 3). This can be accomplished by deprotonation of
the phenolic OH group with P2-Et base (1 equiv), which is
defined as input In1. It thus results that only one input order
(In1 followed by In2) yields the SP form and its correspond-
ing high fluorescence output signal. Application of In2 and
then In1 leaves the fluorescence output low. Because of the
efficient EET fluorescence quenching in the ME and MEH+
forms of the dyads, the binary input combinations 00, 01 and
10 also produce low-output signals (see Figure 4 for the ex-
ample of 2-SP). In essence, this corresponds to a two-input
Figure 3. Streak-camera fluorescence decays (lexc =398 nm; averaged for
lobs =500–540 nm) for 3-M (filled circles), 3-SP (filled squares), 3-ME
(empty squares) and 3-MEH+ (empty circles). The solid lines are to
guide the eye, but do not correspond to the actual fits. Note that the
model 3-M shows essentially no decay on the timescale considered. The
decay curve of 3-ME contains still considerable contributions from the
closed 3-SP (photostationary state). The decay of 3-MEH+ contains a
higher proportion of the short-lived component than that of 3-ME. This
is probably the result of suppression of thermal closing of 3-MEH+
during the course of the measurement.
6496
www.chemeurj.org
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 6492 – 6500

Molecular Implementation of Sequential and Reversible Logic
FULL PAPER
inputs would have to be re-
moved by addition of competi-
tive ligands, which could poten-
tially interfere in the repeated
device operation.
Achievement of reversible mo-
lecular logic through reading of
two compartmentalized logic
gates: Reversible logic is based
on unambiguous mapping of
input vectors to output vectors
and vice versa. Hence, in this
logic scheme each binary input
combination leads to a unique
output combination. A visibly
different fluorescence output
for each input combination is a
promising approach for photon-
ic coding of multi-input events.
On the other hand, any basic
two-input logic gate, possessing
one output channel, is irreversi-
ble and information is lost upon
processing the inputs. To ach-
Scheme 3. Switching of 2-MEH+ (initial state) by sequential application of inputs In1 and In2. Reversible logic
ieve reversibility,
a
second
by combinational switching of 2-MEH+ (initial state) by application of In1 and In3 and of 5H+ by application
of In1.
output that should enable ex-
Figure 4. Sequential logic with 2-MEH+ as initial state and In1 (phospha-
zene base) and In2 (irradiation at l>530 nm) as inputs. The horizontal
line shows the threshold. Fluorescence signals above the threshold are re-
garded as binary 1 and below as binary 0.
Figure 5. Repeated switching cycles of 2-SP. The application order of UV
light (302 nm), acid (TFA), base (P2-Et) and light of l>530 nm is shown
for the first cycle and is identical for the subsequent cycles.
priority AND gate (2-PAND).[29] The switching can be re-
peated for at least five consecutive cycles with an acceptable
dynamic OFF/ON range (Figure 5). According to the inter-
conversions shown in Scheme 1, resetting to the initial state
(MEH+ form) is possible either through application of UV
light and addition of trifluoroacetic acid TFA (SP!MEH+)
or solely through addition of TFA (ME!MEH+). It is
noteworthy that acid–base neutralizations are complete and
fast reactions, producing just water and salts as byproducts.
This is in contrast to some previously reported molecular
keypad lock systems,[35,37,38] in which metal ion or anion
actly one non-repeated output combination per input vector
is needed. It should be noted that the term “reversible
logic” is not to be confused with “chemical reversibility”,
which is a precondition for recycling and resetting of a logic
device (see, for example, Figure 5 for the case discussed
above). In accordance with the necessity of two output
channels, a PET-based anthracene fluorescence switch (com-
pound 5) and the spiropyran–fluorophore dyad 2-SP were
used. Each shows fluorescence in spectrally well-distinguish-
Chem. Eur. J. 2011, 17, 6492 – 6500
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6497

U. Pischel, J. Andrꢁasson et al.
ed regions and was operated in a separate cuvette. This
compartmentalization approach has been used previously by
others for the achievement of logic networks.[54–56,58]
As the initial states, compounds 5H+ and 2-MEH+
(Scheme 3) were chosen, and the inputs (Figure 6) consisted
of the addition of P2-Et base (1 equiv, In1) and irradiation
with light of l>420 nm (In3). The outputs are defined as O1
(aminonaphthalimide fluorescence) and O2 (anthracene
fluorescence). Irradiation of 2-MEH+ (In3) converts this
Scheme 4. Representation of the logic circuit corresponding to the logi-
cally reversible switching of the 2-MEH+/5-H+ combination by In1 and
In3.
Table 2. Truth table of the reversible logic gate 5H+/2-MEH+.
[a]
[b]
[c]
In1
In3
O1 (If,rel
)
O2[d] (If,rel
)
0
1
0
1
0
0
1
1
0 (0.62)
0 (0.64)
1 (1.00)
1 (1.00)
1 (0.99)
0 (0.03)
1 (1.00)
0 (0.03)
[a] P2-Et (1 equiv). [b] l>420 nm, 210 s. [c] Aminonaphthalimide fluores-
cence, lobs =512 nm; threshold
lobs =416 nm; threshold If,rel =0.7.
If,rel =0.7. [d] Anthracene fluorescence,
Figure 6. Normalized fluorescence spectra (lexc =375 nm) of 5H+ (left)
and 2-MEH+ (right) upon addition of base (In1) and upon irradiation at
l>420 nm (In3). The input that actually leads to effective fluorescence
switching is indicated for each system. The other input for the corre-
sponding system (In3 for 5H+ and In1 for 2-MEH+) does not change the
fluorescence. The dotted vertical lines mark the readout wavelengths.
corresponds exactly to one unique output combination. The
chemical processes (light-induced or acid–base-triggered)
for 2-SP/2-MEH+ are reminiscent of the data shown above
(Figure 5; the selective irradiation of ME at l>530 nm is
substituted by light of l>420 nm, which isomerizes both the
MEH+ and the ME forms to the SP form). Hence, the recy-
cling and reset capabilities of the system are identical. The
recycling of the anthracene switch 5 for various alternating
additions of acid and base is also possible (see the Support-
ing Information).
dyad to its 2-SP form, which shows the highest fluorescence
output. Deprotonation (application of In1) yields 2-ME, in
which EET still keeps the fluorescence output O1 low. Ap-
plication of both inputs (In1 and In3), regardless of their
order, again yields the 2-SP form (O1 high). It is important
to note that a change in the input irradiation wavelength to
l>420 nm, at which both the ME form and the MEH+
form are simultaneously irradiated, removes the above de-
signed sequential feature and leads to combinational logic.
Indeed, no sequential logic is required for the design of the
reversible logic device. The behaviour of 2-MEH+ conforms
to a TRANSFER gate (see upper logic circuit with In1, In3,
O1 in Scheme 4).[59] On the other hand, deprotonation
through addition of base (In1) to 5H+ as the initial state of
the other switch activates PET from the appended amino
function, yielding practically quantitative fluorescence
quenching (>95%). The application of In3 has no conse-
quence for the anthracene switch, which lacks significant ab-
sorption above 400 nm. This is an important point, because
the direct irradiation of anthracene could yield photoprod-
ucts, which is not desired in the present context of reversible
logic. In this scenario 5H+ acts as an INVERTER gate with
respect to In1 (or to a more complex circuit with respect to
both inputs In1 and In3; see lower part with output O2 of the
circuit in Scheme 4). The combination of both switches in
one truth table (Table 2) shows that each input combination
Conclusion
The emissions of carefully selected fluorophores (amino-
naphthalimide, dansyl and perylene) appended to a spiro-
pyran can be switched by control of electronic energy trans-
fer. These dyads were designed to yield large spectral over-
lap between the fluorophore emission and the absorption of
the open merocyanine forms of the photochrome, whereas
the closed spiropyran form has a zero overlap integral. The
dynamic range of the fluorescence switching is dependent
on the photostationary state distribution of the spiro form
and the merocyanine ME, which in all cases reached an ap-
proximate ratio of 55:45 (%SP/%ME). This corresponds to
an OFF/ON switching factor of about 2. The fluorescence
modulation can be used in the design of a sequential molec-
ular logic device, working as a two-input priority AND gate
(2-PAND), and in conjunction with an acidochromic anthra-
cene PET switch a logically reversible gate resulted. Only a
few examples of these two comprehensive logic functions
6498
www.chemeurj.org
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 6492 – 6500

Molecular Implementation of Sequential and Reversible Logic
FULL PAPER
Removal of the solvent with a rotary evaporator yielded the crude prod-
uct, which was purified by flash chromatography (SiO2) with a mixture of
dichloromethane/methanol (95:5) as eluent to afford 2-SP (34 mg, 54%).
1H NMR (400 MHz, CDCl3, 258C): d=8.57 (dd, J=7.2, 0.8 Hz, 1H; Ar-
H), 8.37 (d, J=8.4 Hz, 1H; Ar-H), 8.10 (dd, J=8.4, 0.8 Hz, 1H; Ar-H),
7.95 (dd, J=8.8, 2.4 Hz, 1H; Ar-H), 7.90 (d, J=2.4 Hz, 1H; Ar-H), 7.63
(t, J=8.0 Hz, 1H; Ar-H), 7.12 (dt, J=8.0, 1.2 Hz, 1H; Ar-H), 7.03 (dd,
J=7.6, 0.8 Hz, 1H; Ar-H), 6.88–6.76 (m, 3H; Ar-H), 6.66 (d, J=8.8 Hz,
1H; Ar-H), 5.58 (d, J=7.6 Hz, 1H; Ar-H), 6.34 (m, 1H; NH), 5.80 (d,
J=10.4 Hz, 1H; vinyl-H), 5.01 (brs, 2H; NH2), 4.40–4.32 (m, 2H; CH2),
3.22–3.06 (m, 2H; CH2), 3.66–3.56 (m, 2H; CH2), 2.22–2.08 (m, 2H;
CH2), 1.94–1.82 (m, 2H; CH2), 1.22 (s, 3H; CH3), 1.12 ppm (s, 3H; CH3);
13C NMR (100.6 MHz, CDCl3, 258C): d=172.7, 165.3, 164.8, 159.8, 150.0,
147.1, 140.9, 135.9, 134.3, 131.9, 130.0, 128.2, 128.0, 127.5, 125.9, 125.0,
122.8, 122.7, 122.1, 121.7, 120.0, 119.6, 118.6, 115.6, 111.2, 109.6, 107.0,
106.9, 52.9, 43.3, 40.2, 39.3, 33.9, 26.1, 24.8, 20.0 ppm; HRMS (TOF,
EI+): m/z: calcd for C36H33N5O6: 631.2431; found: 631.2422.
have been achieved previously. The hybrid approach chosen
here (combination of chemical and photonic signals) is a
promising strategy for achieving high levels of logic func-
tionality. The coinciding photonic natures of the output and
one input should in principle allow concatenation. Finally,
from a conceptual point of view, the chemical input serves
as an example of a sensing input, whereas the photonic
input opens the possibility for remote control of the device.
Experimental Section
Irradiation: Light with l>420 nm and with l>530 nm was generated
with a Xe lamp (150 W) together with long-pass glass filters (l cut-on=
420 nm and 530 nm). The resulting light power densities on the samples
were ꢂ30 mWcmꢀ2 and ꢂ18 mWcmꢀ2, respectively. The 302 nm UV
light was generated with a UVP handheld UV lamp (Model UVM-57,
1.5 mWcmꢀ2 power density). The irradiation time used for the photo-iso-
merisations was 210 s in all cases.
Acknowledgements
Photophysical measurements: All measurements were performed in air-
equilibrated acetonitrile solutions at room temperature. The UV/Vis ab-
sorption spectra were recorded with a UV-1603 spectrophotometer from
Shimadzu or with a CARY 5000 UV/Vis/NIR spectrophotometer. The
fluorescence spectra were recorded with a Cary Eclipse fluorimeter from
Varian. The fluorescence quantum yields were determined with quinine
sulfate (Ff =0.55 in H2SO4 (0.05m))[60] as standard and corrected for dif-
ferences in the refractive indexes of the reference and sample solvents.
The quantum yield of dansyl fluorescence in the dyad 3-SP is corrected
for the absorbance of the spiropyran at the excitation wavelength
(341 nm).
Financial support by the Ministerio de Ciencia e Innovaciꢀn, Madrid
(grant CTQ2008–06777-C02–02/BQU, PhD fellowship for P.R.), the Con-
sejerꢄa de Economꢄa, Innovaciꢀn y Ciencia, Seville (grant P08-FQM-
3685), the Swedish Research Council (grant 622–2010–280), and the Eu-
ropean Research Council (ERC FP7/2007–2013 No. 203952) is gratefully
acknowledged.
[1] A. P. de Silva, S. Uchiyama, Nat. Nanotechnol. 2007, 2, 399–410.
[2] K. Szaciłowski, Chem. Rev. 2008, 108, 3481–3548.
[3] J. Andrꢁasson, U. Pischel, Chem. Soc. Rev. 2010, 39, 174–188.
[4] U. Pischel, Aust. J. Chem. 2010, 63, 148–164.
[5] A. Credi, Angew. Chem. 2007, 119, 5568–5572; Angew. Chem. Int.
Ed. 2007, 46, 5472–5475.
The absorption spectra shown in Figure 1 were recorded with samples in
which virtually 100% had been converted to the appropriate forms. The
experimental protocol has been described in the literature previously.[61]
The fluorescence lifetimes of 2-M, 3-M and 4-M and the corresponding
lifetimes of 2-SP, 3-SP and 4-SP were measured by use of pulsed laser
diodes at 377 nm or 405 nm (PicoQuant) with 10 MHz repetition rates as
the excitation sources. The emitted photons were recorded at 505 nm
under magic angle conditions and were detected with a thermoelectrical-
ly cooled microchannel plate photomultiplier tube (MCP-PMT R3809U-
50; Hamamatsu). The fluorescence signals were recorded until 10000
counts were obtained in the top channel.
[6] S. Angelos, Y.-W. Yang, N. M. Khashab, J. F. Stoddart, J. I. Zink, J.
Am. Chem. Soc. 2009, 131, 11344–11346.
[7] R. J. Amir, M. Popkov, R. A. Lerner, C. F. Barbas III, D. Shabat,
Angew. Chem. 2005, 117, 4452–4455; Angew. Chem. Int. Ed. 2005,
44, 4378–4381.
[8] S. Ozlem, E. U. Akkaya, J. Am. Chem. Soc. 2009, 131, 48–49.
[9] A. P. de Silva, M. R. James, B. O. F. McKinney, D. A. Pears, S. M.
Weir, Nat. Mater. 2006, 5, 787–790.
The lifetimes of the ME and MEH+ forms of the dyads (2-ME, 3-ME, 4-
ME, 2-MEH+, 3-MEH+, 4-MEH+) were measured with a streak camera
system. The excitation pulses were provided by a Tsunami Ti:sapphire
laser (Spectra-Physics) that was pumped with a Millennia Pro X laser
(Spectra-Physics). The Tsunami laser output was tuned to 796 nm and
subsequently frequency-doubled to 398 nm. The photons emitted from
the samples were passed through a spectrograph (Acton SP2300, Prince-
ton Instruments) and were registered with a streak camera (C5680, Ha-
mamatsu) with a synchroscan unit (M5675, Hamamatsu). Every single
frame was measured and stored individually and the time-resolved fluo-
rescence spectra were obtained after jitter correction. From these time-
resolved fluorescence spectra, the fluorescence time profiles were ex-
tracted and fitted to biexponential expressions, including the instrument
response function. The samples of the ME and the MEH+ forms of the
dyads were repeatedly refreshed with doses of 302 nm light (1 min)
during the time-resolved fluorescence measurements.
[10] D. C. Magri, G. J. Brown, G. D. McClean, A. P. de Silva, J. Am.
Chem. Soc. 2006, 128, 4950–4951.
[11] D. Margulies, A. D. Hamilton, J. Am. Chem. Soc. 2009, 131, 9142–
9143.
[12] T. Konry, D. R. Walt, J. Am. Chem. Soc. 2009, 131, 13232–13233.
[13] O. A. Bozdemir, R. Guliyev, O. Buyukcakir, S. Selcuk, S. Kolemen,
G. Gulseren, T. Nalbantoglu, H. Boyaci, E. U. Akkaya, J. Am.
Chem. Soc. 2010, 132, 8029–8036.
[14] F. M. Raymo, M. Tomasulo, Chem. Soc. Rev. 2005, 34, 327–336.
[15] F. M. Raymo, M. Tomasulo, J. Phys. Chem. A 2005, 109, 7343–7352.
[16] S. L. Gilat, S. H. Kawai, J. M. Lehn, Chem. Eur. J. 1995, 1, 275–284.
[17] A. Fernꢅndez-Acebes, J.-M. Lehn, Chem. Eur. J. 1999, 5, 3285–
3292.
[18] M. Irie, Chem. Rev. 2000, 100, 1685–1716.
[19] T. B. Norsten, N. R. Branda, Adv. Mater. 2001, 13, 347–349.
[20] A. J. Myles, N. R. Branda, Adv. Funct. Mater. 2002, 12, 167–173.
[21] T. Fukaminato, T. Sasaki, T. Kawai, N. Tamai, M. Irie, J. Am. Chem.
Soc. 2004, 126, 14843–14849.
[22] M. Berberich, A.-M. Krause, M. Orlandi, F. Scandola, F. Wꢆrthner,
Angew. Chem. 2008, 120, 6718–6721; Angew. Chem. Int. Ed. 2008,
47, 6616–6619.
[23] J. Kꢇrnbratt, M. Hammarson, S. Li, H. L. Anderson, B. Albinsson, J.
Andrꢁasson, Angew. Chem. 2010, 122, 1898–1901; Angew. Chem.
Int. Ed. 2010, 49, 1854–1857.
Representative procedure for the synthesis of the fluorophore-spiropyran
dyads (compound 2-SP): A solution of the succinimidyl ester of 1-SP
(50 mg, 0.1 mmol) and N-(2-aminoethyl)-4-amino-1,8-naphthalimide
(26 mg, 0.1 mmol) in tetrahydrofuran/dichloromethane (2:3, 8 mL) and
N,N-dimethylformamide (4 mL) was stirred at room temperature for
2.5 h. The resulting mixture was diluted with dichloromethane and
washed with brine and water. The organic phase was dried over Na2SO4.
Chem. Eur. J. 2011, 17, 6492 – 6500
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6499

U. Pischel, J. Andrꢁasson et al.
[24] J. Andrꢁasson, S. D. Straight, S. Bandyopadhyay, R. H. Mitchell,
T. A. Moore, A. L. Moore, D. Gust, Angew. Chem. 2007, 119, 976–
979; Angew. Chem. Int. Ed. 2007, 46, 958–961.
[25] J. Andrꢁasson, S. D. Straight, S. Bandyopadhyay, R. H. Mitchell,
T. A. Moore, A. L. Moore, D. Gust, J. Phys. Chem. C 2007, 111,
14274–14278.
[26] J. Andrꢁasson, S. D. Straight, T. A. Moore, A. L. Moore, D. Gust, J.
Am. Chem. Soc. 2008, 130, 11122–11128.
[27] J. Andrꢁasson, G. Kodis, Y. Terazono, P. A. Liddell, S. Bandyopad-
hyay, R. H. Mitchell, T. A. Moore, A. L. Moore, D. Gust, J. Am.
Chem. Soc. 2004, 126, 15926–15927.
ref. [41]); b) aminonaphthalimide: Eox =1.20 V, Ered =ꢀ1.41 V (both
in butyronitrile, ref. [44]), E*=2.7 eV (determined for 2M according
to ref. [44]); c) dansyl: Eox =0.9 V, Ered =ꢀ2.02 V (both in acetoni-
trile, ref. [45]), E*=3.0 eV (determined for 3M according to
ref. [44]); d) perylene: Eox =0.85 V (in acetonitrile, ref. [46]), Ered
=
ꢀ1.67 V (in N,N-dimethylformamide, ref. [46]), E*=2.8 eV (deter-
mined for 4M from the intersection of the normalized absorption
and emission spectra). The redox potentials refer to closely related
model compounds as used in the cited references. The Coulomb
term C was taken as ꢀ0.06 eV in acetonitrile.
[44] S. R. Greenfield, W. A. Svec, D. Gosztola, M. R. Wasielewski, J.
Am. Chem. Soc. 1996, 118, 6767–6777.
[28] J. Andrꢁasson, S. D. Straight, G. Kodis, C.-D. Park, M. Hambourger,
M. Gervaldo, B. Albinsson, T. A. Moore, A. L. Moore, D. Gust, J.
Am. Chem. Soc. 2006, 128, 16259–16265.
[45] P. Ceroni, I. Laghi, M. Maestri, V. Balzani, S. Gestermann, M.
Gorka, F. Vçgtle, New J. Chem. 2002, 26, 66–75.
[29] J. Andrꢁasson, S. D. Straight, T. A. Moore, A. L. Moore, D. Gust,
Chem. Eur. J. 2009, 15, 3936–3939.
[46] M. Montalti, A. Credi, L. Prodi, M. T. Gandolfi, Handbook of Pho-
tochemistry, CRC Press, Boca Raton, 2006.
[30] J. L. Bahr, G. Kodis, L. de la Garza, S. Lin, A. L. Moore, T. A.
Moore, D. Gust, J. Am. Chem. Soc. 2001, 123, 7124–7133.
[31] M. Tomasulo, E. Deniz, R. J. Alvarado, F. M. Raymo, J. Phys. Chem.
C 2008, 112, 8038–8045.
[32] L. Y. Zhu, W. W. Wu, M. Q. Zhu, J. J. Han, J. K. Hurst, A. D. Q. Li,
J. Am. Chem. Soc. 2007, 129, 3524–3526.
[33] S. Mao, R. K. P. Benninger, Y. L. Yan, C. Petchprayoon, D. Jackson,
C. J. Easley, D. W. Piston, G. Marriott, Biophys. J. 2008, 94, 4515–
4524.
[47] A. P. de Silva, H. Q. N. Gunaratne, J.-L. Habib-Jiwan, C. P. McCoy,
T. E. Rice, J.-P. Soumillion, Angew. Chem. 1995, 107, 1889–1891;
Angew. Chem. Int. Ed. Engl. 1995, 34, 1728–1731.
[48] Y. Q. Gao, R. A. Marcus, J. Phys. Chem. A 2002, 106, 1956–1960.
[49] G. de Ruiter, E. Tartakovsky, N. Oded, M. E. van der Boom, Angew.
Chem. 2010, 122, 173–176; Angew. Chem. Int. Ed. 2010, 49, 169–
172.
[50] U. Pischel, Angew. Chem. 2010, 122, 1396–1398; Angew. Chem. Int.
Ed. 2010, 49, 1356–1358.
[34] Z. Guo, W. Zhu, L. Shen, H. Tian, Angew. Chem. 2007, 119, 5645–
5649; Angew. Chem. Int. Ed. 2007, 46, 5549–5553.
[35] D. Margulies, C. E. Felder, G. Melman, A. Shanzer, J. Am. Chem.
Soc. 2007, 129, 347–354.
[36] W. Sun, C. Zhou, C.-H. Xu, C.-J. Fang, C. Zhang, Z.-X. Li, C.-H.
Yan, Chem. Eur. J. 2008, 14, 6342–6351.
[37] M. Suresh, A. Ghosh, A. Das, Chem. Commun. 2008, 3906–3908.
[38] S. Kumar, V. Luxami, R. Saini, D. Kaur, Chem. Commun. 2009,
3044–3046.
[51] U. Pischel, J. Andrꢁasson, New J. Chem. 2010, 34, 2701–2703.
[52] F. M. Raymo, R. J. Alvarado, S. Giordani, M. A. Cejas, J. Am.
Chem. Soc. 2003, 125, 2361–2364.
[53] F. M. Raymo, S. Giordani, Org. Lett. 2001, 3, 3475–3478.
[54] F. M. Raymo, S. Giordani, Org. Lett. 2001, 3, 1833–1836.
[55] F. M. Raymo, S. Giordani, J. Am. Chem. Soc. 2002, 124, 2004–2007.
[56] F. M. Raymo, S. Giordani, Proc. Natl. Acad. Sci. USA 2002, 99,
4941–4944.
[57] S. Silvi, E. C. Constable, C. E. Housecroft, J. E. Beves, E. L.
Dunphy, M. Tomasulo, F. M. Raymo, A. Credi, Chem. Eur. J. 2009,
15, 178–185.
[39] P. Remꢀn, R. Ferreira, J.-M. Montenegro, R. Suau, E. Pꢁrez-Inestro-
sa, U. Pischel, ChemPhysChem 2009, 10, 2004–2007.
[40] M. Semeraro, A. Credi, J. Phys. Chem. C 2010, 114, 3209–3214.
[41] A. A. Garcꢄa, S. Cherian, J. Park, D. Gust, F. Jahnke, R. Rosario, J.
Phys. Chem. A 2000, 104, 6103–6107.
[42] A. Garcꢄa, M. Marquꢁz, T. Cai, R. Rosario, Z. Hu, D. Gust, M.
Hayes, S. A. Vail, C.-D. Park, Langmuir 2007, 23, 224–229.
[43] The following electrochemical (all vs. SCE in polar solvents) and
spectroscopic data were used for the calculation of the free energy
changes for oxidative and reductive electron transfer in the dyads:
[58] K. Szaciłowski, Chem. Eur. J. 2004, 10, 2520–2528.
[59] H. T. Baytekin, E. U. Akkaya, Org. Lett. 2000, 2, 1725–1727.
[60] J. V. Morris, M. A. Mahaney, J. R. Huber, J. Phys. Chem. 1976, 80,
969–974.
[61] T. Stafforst, D. Hilvert, Chem. Commun. 2009, 287–288.
Received: January 4, 2011
a) spiropyran:
E
ox =1.20 V,
E
red =ꢀ1.25 V (both in benzonitrile,
Published online: April 27, 2011
6500
www.chemeurj.org
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 6492 – 6500