Molecular all-photonic encoder-decoder

Article abstract of DOI:10.1021/ja802845z

In data processing, an encoder can compress digital information for transmission or storage, whereas a decoder recovers the information in its original form. We report a molecular triad consisting of a dithienylethene covalently linked to two fulgimide ph

Full text of DOI:10.1021/ja802845z

Published on Web 07/29/2008
Molecular All-Photonic Encoder-Decoder
Joakim Andre´asson,*,† Stephen D. Straight,‡ Thomas A. Moore,*,‡ Ana L. Moore,*,‡
and Devens Gust*,‡
Department of Chemical and Biological Engineering, Chalmers UniVersity of Technology,
SE-412 96 Go¨teborg, Sweden, and Department of Chemistry and Biochemistry,
Arizona State UniVersity, Tempe, Arizona 85287-1604
Received April 17, 2008; E-mail: a-son@chalmers.se; gust@asu.edu; tom.moore@asu.edu; amoore@asu.edu
Abstract: In data processing, an encoder can compress digital information for transmission or storage,
whereas a decoder recovers the information in its original form. We report a molecular triad consisting of
a dithienylethene covalently linked to two fulgimide photochromes that performs as an all-photonic single-
bit 4-to-2 encoder and 2-to-4 decoder. The encoder compresses the information contained in the four inputs
into two outputs. The inputs are light of four different wavelengths that photoisomerize the fulgimide,
dithienylethene, or both. The outputs are absorbance at two wavelengths. The two decoder inputs are
excitation at two wavelengths, whereas the four outputs, which recover the information compressed into
the inputs, are absorbance at two wavelengths, transmittance at one wavelength, and fluorescence emission.
The molecule can be cycled through numerous encoder and decoder functions without significant
photodecomposition. Molecular photonic encoders and decoders could potentially be used for labeling and
tracking of nano- and microscale objects as well as for data manipulation.
Table 1. Truth Table for the 4-to-2 Encoder
Introduction
input I0
input I1
input I2
input I3
output O1
output O0
It is becoming well-established that because some molecules
can act as reversible switches in response to stimuli, they can
be used to perform digital data-processing functions, much as
is done by the transistor switches in conventional electronics.
Molecular switches, Boolean logic gates, and other data
manipulation devices have been described.1–11 Although most
of the initial reports of such systems were of simple switches
or single-function logic gates, more recent work has produced
molecules or mixtures of molecules capable of performing a
variety of logic functions.12–19 Most often, these systems use
1
0
0
0
0
1
0
0
0
0
1
0
0
0
0
1
0
0
1
1
0
1
0
1
one or more chemical inputs. We have been investigating the
use of light as inputs and outputs, usually via photochromes
that interact photochemically or photophysically with attached
chromophores. Photochromes are photochemical switches that
may be photoisomerized back and forth between two (meta)-
stable states.20 Recent implementations include a half-adder,21
2:1 digital multiplexer,22 and 1:2 demultiplexer.23
A digital encoder converts data into a code. One reason to
do this is to compress information for transmission or storage.
Consider, for example, the single-bit 4-to-2 encoder, whose truth
table is shown in Table 1. The encoder converts 4 bits of data
† Chalmers University of Technology.
‡ Arizona State University.
(1) Ball, P. Nature (London, U.K.) 2000, 406, 118–120.
(2) Magri, D. C.; Vance, T. P.; de Silva, A. P. Inorg. Chim. Acta 2007,
360, 751–764.
(3) Tian, H.; Wang, Q. C. Chem. Soc. ReV. 2006, 35, 361–374.
(4) Venturi, M.; Balzani, V.; Ballardini, R.; Credi, A.; Gandolfi, M. T.
Int. J. Photoenergy 2004, 6, 1–10.
(5) de Silva, A. P.; Uchiyama, S. Nat. Nanotechnol. 2007, 2, 399–410.
(6) de Silva, A. P.; Uchiyama, S.; Vance, T. P.; Wannalerse, B. Coord.
Chem. ReV. 2007, 251, 1623–1632.
(17) Frezza, B. M.; Cockroft, S. L.; Ghadiri, M. R. J. Am. Chem. Soc.
2007, 129, 14875–14879.
(7) Balzani, V.; Credi, A.; Venturi, M. ChemPhysChem 2003, 3, 49–59.
(8) Credi, A. Angew. Chem., Int. Ed. 2007, 46, 5472–5475.
(9) Brown, G. J.; de Silva, A. P.; Pagliari, S. Chem. Commun. (Cambridge
U.K.) 2002, 2461–2463.
(18) Browne, W. R.; De Jong, J. J. D.; Kudernac, T.; Walko, M.; Lucas,
L. N.; Uchida, K.; van Esch, J. H.; Feringa, B. L. Chem.sEur. J.
2005, 11, 6414–6429.
(19) Strack, G.; Ornatska, M.; Pita, M.; Katz, E. J. Am. Chem. Soc. 2008,
130, 4234–4235.
(10) Pischel, U. Angew. Chem., Int. Ed. 2007, 46, 4026–4040.
(11) Raymo, F. M.; Giordani, S. J. Am. Chem. Soc. 2001, 123, 4651–4652.
(12) Niazov, T.; Baron, R.; Katz, E.; Lioubashevski, O.; Willner, I. Proc.
Natl. Acad. Sci. U.S.A. 2006, 103, 17160–17163.
(20) Gust, D.; Moore, T. A.; Moore, A. L. Chem. Commun. (Cambridge
U.K.) 2006, 2006, 1169–1178.
(21) Andre´asson, J.; Straight, S. D.; Kodis, G.; Park, C.-D.; Hambourger,
M.; Gervaldo, M.; Albinsson, B.; Moore, T. A.; Moore, A. L.; Gust,
D. J. Am. Chem. Soc. 2006, 12, 16259–16265.
(13) Margulies, D.; Melman, G.; Felder, C. E.; Arad-Yellin, R.; Shanzer,
A. J. Am. Chem. Soc. 2004, 126, 15400–15401.
(14) Liu, Y.; Jiang, W.; Zhang, H. Y.; Li, C. J. J. Phys. Chem. B 2006,
110, 14231–14235.
(22) Andre´asson, J.; Straight, S. D.; Bandyopadhyay, S.; Mitchell, R. H.;
Moore, T. A.; Moore, A. L.; Gust, D. Angew. Chem., Int. Ed. 2007,
46, 958–961.
(15) Baron, R.; Lioubashevski, O.; Katz, E.; Niazov, T.; Willner, I. J. Phys.
Chem. A 2006, 110, 8451–8456.
(23) Andre´asson, J.; Straight, S. D.; Bandyopadhyay, S.; Mitchell, R. H.;
Moore, T. A.; Moore, A. L.; Gust, D. J. Phys. Chem. C 2007, 111,
14274–14278.
(16) Margulies, D.; Felder, C. E.; Melman, G.; Shanzer, A. J. Am. Chem.
Soc. 2007, 129, 347–354.
9
11122 J. AM. CHEM. SOC. 2008, 130, 11122–11128
10.1021/ja802845z CCC: $40.75
2008 American Chemical Society

Molecular All-Photonic Encoder-Decoder
A R T I C L E S
Figure 2. Structures of fulgimide (2) and dithienylethene (3) model
compounds in the open (left) and closed (right) forms. Although the
molecules used in this study were racemic, only one enantiomer is shown
in the figure.
Figure 1. Structures of triad encoder/decoder 1 with all photochromes in
the open states that absorb only UV light (a) and in the closed, cyclic states
that absorb in the visible (b). The designations FG and DTE indicate
fulgimide and dithienylethene moieties, respectively. Although the molecules
used in this study were mixtures of racemic compounds, only one enantiomer
is shown in the figure.
Figure 3. Absorption spectra of 2-methyltetrahydrofuran solutions of model
compounds 2o (O), 3o (9), 2c (-O-), and 3c (-9-). The spectra have
been scaled to represent the relative populations of chromophores in 1.
into 2 bits. Each of the four inputs, I0-I3, can assume a value
of either 0 or 1. The two outputs, O0 and O1 comprise a 2-bit
binary number (0, 1, 2, or 3 in base-10 mathematics), where
O0 is the sum digit and O1 is the carry digit. A decoder converts
the compressed data back into the original form.
Model Compound Spectroscopy. Model FG 2 and DTE 3
(Figure 2) have been previously reported.24 Fulgimide 2o, in
the open form, exists as a mixture of E- and Z-isomers that
absorb maximally at 380 nm in 2-methyltetrahydrofuran (Figure
3). Irradiation with UV light (e.g., 397 nm) leads to isomer-
ization between the E- and Z-forms and cyclization of the
E-form, ultimately converting the sample to a photostationary
distribution greatly enriched in the closed form 2c. This isomer
has absorption maxima at 262, 295, and 510 nm). The closed
form 2c fluoresces (λmax ) 600 nm), and single-photon-timing
experiments have determined that the excited singlet state has
a lifetime of 135 ps.25 Irradiation of the closed form with visible
light induces photoisomerization back to 2o. Thermal isomer-
ization is not observed on the time scale of these experiments.
The model dithienylethene in the open form, 3o (Figure 2)
has an absorption maximum at 306 nm, with no absorption in
the visible (Figure 3). Irradiation in the UV (e.g., 302 nm)
isomerizes 3o to the closed, cyclic form 3c, which has absorption
maxima at 274, 335, and 600 nm (Figure 3). Photoisomerization
with visible light converts 3c back to 3o, and thermal reversion
is not observed.
Here, we report molecular triad 1 (Figure 1), which performs
as an all-photonic 4-to-2 encoder and 2-to-4 decoder. The
molecule consists of a dithienylethene (DTE) photochrome
covalently linked to two fulgimide (FG) photochromic moieties.
The two kinds of photochromes may be isomerized indepen-
dently by different wavelengths of light between closed forms
which are colored in the visible (Figure 1b) and open, UV-
absorbing forms (Figure 1a). The DTE and FG also have
absorption maxima at different wavelengths in both the open
and closed forms. The closed form of the fulgimide is
fluorescent, but when the dithienylethene is also in the closed
form, the fulgimide fluorescence is strongly quenched by
intramolecular processes. These properties allow the triad to
act as either a 4-to-2 encoder or a 2-to-4 decoder, depending
upon the nature of the photonic inputs and outputs, as explained
in detail below.
Triad 1 Spectroscopy. Because the triad contains three
photochromes, two of which are identical, it can exist in six
photoisomeric forms. The absorption spectra of 2-methyltet-
rahydrofuran solutions highly enriched in each of the isomers
of interest for this investigation are shown in Figure 4.
Results and Discussion
The synthesis and characterization of triad 1 and its precursors
and the details of the spectroscopic measurements are given in
the Experimental Section. Here, we will first discuss the
photochemistry of two model compounds and then that of the
triad. Finally, we will show how these photochemical properties
can be employed to achieve the encoder and decoder functions.
(24) Straight, S. D.; Liddell, P. A.; Terazono, Y.; Moore, T. A.; Moore,
A. L.; Gust, D. AdV. Funct. Mater. 2007, 17, 777–785.
(25) Straight, S. D.; Terazono, Y.; Kodis, G.; Moore, T. A.; Moore, A. L.;
Gust, D. Aust. J. Chem. 2006, 59, 170–174.
9
J. AM. CHEM. SOC. VOL. 130, NO. 33, 2008 11123

A R T I C L E S
Andre´asson et al.
Table 2. Time Constants (τ) for Photoisomerization Reactions
compd
λ (nm) τ (s)
compd
λ (nm)
τ (s)
2o f 2c 397a
2o f 2c 366b
120 1, FGo-DTE f FGc-DTE
30 1, FGo-DTE f FGc-DTE
397a
170
6, 100c
366b
2c f 2o greend 21 1, FGc-DTEo f FGo-DTEo greend 22
3o f 3c 302e
3o f 3c 366b
4
1, FGc-DTEc f FGo-DTE greend 280
47 1, FG-DTEo f FG-DTEc
302e
366b
4
25
3c f 3o greend 350 1, FG-DTEo f FG-DTEc
1, FG-DTEc f FG-DTEo
greend 430
a 1.6 mW/cm2, see the Experimental Section for details. b 1.5 mW/
cm2, see the Experimental Section for details. c The decay was
biexponential under these conditions. d 9 mW/cm2, 460 < λ < 590 nm,
see the Experimental Section for details. e 1.5 mW/cm2, see the
Experimental Section for details.
Figure 4. Absorption spectra of a 2-methytetrahydrofuran solution of triad
1 after irradiation with light to produce solutions very highly enriched in
FGo-DTEo (s), FGc-DTEo (O), FGo-DTEc (9), and FGc-DTEc
(- - -). Also shown is the fluorescence emission at 630 nm from the
fulgimide in FGc-DTEo excited at 500 nm (0).
Table 3. Inputs and Outputs for Encoder Function
inputs (I) and
outputs (O)
λ (nm)
conditions
Irradiation of the solution with broad-band green light (460 <
λ < 590 nm) yields the triad in which all three photochromes
are in the open forms, which do not absorb in the visible.
Absorption maxima are observed at 286 (sh), 331, and 387 (sh)
nm. This form of 1 will be abbreviated FGo-DTEo, and we
will make no distinction between the two equivalent fulgimides.
Two fulgimides, rather than one, were incorporated into the
molecule in order to enhance some of the effects described
below. Irradiation of the solution at 302 nm, where most of the
absorption is due to the DTEo chromophore, yields a solution
greatly enriched in FGo-DTEc, which has maxima at 362 and
604 nm. Alternatively, irradiating FGo-DTEo at 397 nm
photoisomerizes the fulgimides to the closed form, giving
FGc-DTEo. Initially, of course, this irradiation produces
molecules in which one fulgimide is in the closed form and
one is still open, but continued irradiation yields samples in
which most of the fulgimides are closed in the photostationary
distribution. This material has absorption maxima at 270, 324,
and 511 nm (Figure 4). The figure also shows the fulgimide
fluorescence emission spectrum of FGc-DTEo, obtained with
500 nm excitation. The emission maximum is at 630 nm.
Finally, if FGo-DTEo is irradiated with both 302 and 397 nm
light, the sample is converted to a solution highly enriched in
FGc-DTEc. The absorption maxima are at 266, 357, 535, and
598 nm.
Figure 4 demonstrates both that it is possible to obtain
photostationary distributions very highly enriched in the four
photoisomers mentioned above and that the spectra of the triad
in its various forms are very similar to linear combinations of
the spectra of the model compounds shown in Figure 3. That
is, linking the chromophores covalently has not led to strong
electronic interactions that influence the absorption spectra.
In the photostationary distribution containing mainly the
FGc-DTEc form of the triad, the FGc fluorescence that was
apparent in the FGc-DTEo form (Figure 4) and in model
compound 2c is quenched by 98%. This quenching does not
lead to sample degradation or the formation of stable photo-
products. It is most likely due to singlet-singlet energy transfer
from FGc to DTEc. The overlap of FGc fluorescence and DTEc
absorption is excellent, and the interchromophore separations
are such that energy transfer by the Fo¨rster mechanism is
expected to be very rapid. Because DTEc is essentially
nonfluorescent, energy transfer could not be confirmed by
fluorescence excitation spectroscopy. Calculations using Fo¨rster
theory (ΦF of FGc ) 0.005, ε600 of DTEc ) 2 × 104 M-1
cm-1, orientation factor κ2 ) 0.67) yield a value for the critical
distance R0 of 23 Å. Molecular mechanics modeling of 1 yields
I0 and reset
I1
I2
I3
O0
O1
460 < λ < 590 nm
9 mW/cm2, 30 min
1.6 mW/cm2, 4 min
1.5 mW/cm2, 3 s
1.5 mW/cm2, 45 s
absorbance
397
302
366
475
625
absorbance
a center-to-center distance between the chromophores of 13 Å.
The resulting predicted energy transfer efficiency is 97%, which
is comparable to that observed experimentally. The Fo¨rster
calculation predicts that the 135 ps lifetime of the FGc first
excited singlet state would be quenched to <4 ps by energy
transfer to DTEc. Time-resolved measurements of this lifetime
using the single-photon-timing technique yield a lifetime of <5
ps, although the accuracy of this measurement is not high
because the instrument response function with which it is
convoluted is 30 ps. Photoinduced electron transfer to yield a
short-lived charge-separated state is an alternative mechanism.
Because the redox potentials of the two moieties are unknown,
this mechanism cannot be ruled out.
Isomerization Kinetics. The observed rates for the various
photoisomerization reactions in the triad and model compounds
depend on the light flux. (The actual photochemical elementary
processes typically occur on the subnanosecond time scale.) The
time constants for selected processes under the conditions used
in the encoder-decoder study are presented in Table 2, along
with the conditions. All studies were carried out in 2-meth-
yltetrahydrofuran at ambient temperatures. In general, the time
constants for isomerization of the chromophores in the triad
are roughly similar to those for the corresponding monomers.
An exception is the time constant for conversion of FGc to FGo
using green light. In model compound 2c and when the DTE
moiety of the triad is in the open form, this time constant is
∼22 s. However, when the DTE in the triad is in the closed
form, the time constant for FGc to FGo increases to 280 s. The
reason is the strong quenching of the FGc first excited singlet
state by DTEc, which is also responsible for the strong
fluorescence quenching mentioned above.
Encoder Function. As mentioned above, a solution of triad
1 in 2-methyltetrahydrofuran functions as a 4-to-2 digital
encoder. The initial state, prior to applying any inputs, is
FGo-DTEo, where all of the photochromes are in the open
forms. This is achieved by irradiation with green light. The triad
is reset to this state prior to each input operation. The various
inputs and outputs are summarized in Table 3. Applying input
I0 does not lead to any isomerization, and the sample remains
in the FGo-DTEo form.
9
11124 J. AM. CHEM. SOC. VOL. 130, NO. 33, 2008

Molecular All-Photonic Encoder-Decoder
A R T I C L E S
Figure 6. Photocycling of triad 1 in the encoder mode. The absorbance at
475 nm represents output O0, and the absorbance at 625 nm represents output
O1. The black bars show the absorbance following repetitive application of
inputs I0, (0), I1 (1), I2 (2), and I3 (3) for five complete cycles. It is apparent
that there is no significant degradation of the sample.
Figure 5. Performance of the triad as a 4-to-2 encoder. The black bars
show the absorbance at the appropriate monitoring wavelengths for outputs
O0 and O1 following application of the four inputs I. The dashed lines are
thresholds used to distinguish an on output from an off output. The variation
in height at the top of each bar (barely visible) shows the variation in
absorbance during measurement for 5 s.
Table 4. Truth Table for the 2-to-4 Decoder
input I1
input I0
output O0
output O1
output O2
output O3
There is no absorption at the two output wavelengths (475
and 625 nm), and both outputs remain in the 0, or off state,
corresponding to 0 in base-10 mathematics. Input I1 isomerizes
the triad to mainly the FGc-DTEo form. Therefore, the
absorbance at 475 nm is high, whereas that at 625 nm is low.
Thus, O0 ) 1, and O1 ) 0 (1 in base-10). Input I2 is absorbed
by the DTE moiety, and converts the triad solution to mainly
FGo-DTEc, giving rise to strong absorption at 625 nm but not
at 475 nm. Output O0 is now 0, and O1 ) 1 (2 in base-10).
Finally, applying I3 at 366 nm isomerizes both photochromes
to the closed forms (FGc-DTEc), giving high absorbance at
both 475 and 625 nm. Both outputs are in the 1 state (on),
corresponding to the number 3 in base-10.
Thus, the triad solution functions as a 4-to-2 encoder,
converting an on state for each of the four inputs (I0-I3) into
the binary representation of a unique base-10 number 0-3 and
compressing four bits of data to two, as required by Table 1.
The actual performance of the molecule is shown in Figure 5,
where Figure 5a shows the absorbance at 475 nm and Figure
5b shows that at 625 nm. The dotted lines in these figures
represent convenient thresholds for determining whether the
output is on or off. Unlike an ideal logic gate, any real device
must function via a threshold value of some kind. The signal-
to-noise obtained is shown by the variation at the tops of the
black bars in Figure 5, each of which represents measurement
of the absorbance for 5 s. The noise is barely visible and far
better than necessary to distinguish on from off.
0
0
1
1
0
1
0
1
1
0
0
0
0
1
0
0
0
0
1
0
0
0
0
1
Table 5. Inputs and Outputs for Decoder Function
inputs (I) and
outputs (O)
λ (nm)
conditions
reset
I0
I1
O0
O1
O2
O3
460 < λ < 590 nm
9 mW/cm2, 30 min
1.6 mW/cm2, 12 min
1.5 mW/cm2, 25 s
transmittance
emission, λex ) 500 nm
absorbance
397
302
535
624
393
535
absorbance
between the inputs. This cycle was repeated five times, with
no significant degradation of the response at either O0 (475 nm)
or O1 (625 nm).
Decoder Function. The 2-to-4 decoder converts binary
information from two coded inputs, such as those generated by
the encoder, to four unique outputs, each of which is either off
(0) or on (1). The truth table is shown in Table 4. Following
application of each input combination 00, 01, 10, or 11, (0, 1,
2, or 3 in base-10), the output corresponding to the base-10
number is switched on. When triad 1 functions as a decoder,
the initial state is FGo-DTEo just as for the encoder. It can be
set from any other state of the device with green light. The inputs
and outputs for the decoder are listed in Table 5.
The performance of 1 in 2-methyltetrahydrofuran as a decoder
is depicted in Figure 7, which shows the response of each output
in the initial state and after applying various combinations of
inputs. When both inputs are off, the sample remains in the
FGo-DTEo form. The absorption spectrum (Figure 4) shows
Photocycling of the Encoder. To have any hope of being part
of a useful device, an encoder must be capable of cycling many
times. Although the focus of this research was not the production
of technologically useful devices, but rather the realization of
scientific principles and possibilities, it was of interest to
investigate the photostability of the triad under the conditions
used for implementing the encoder. Figure 6 shows the
photocycling of the encoder through all four inputs, with resets
9
J. AM. CHEM. SOC. VOL. 130, NO. 33, 2008 11125

A R T I C L E S
Andre´asson et al.
4), giving an on response for O2. Both absorbance and
transmittance at 535 are below the threshold so that O3 and O0
are off. There is no significant emission at 624 nm so that O1 is
off, and the system represents the third line of the truth table.
Finally, applying both I0 and I1 isomerizes both photochromes
so that the triad is in mainly the FGc-DTEc form. Because
FGc absorbs only weakly at 393 nm, the absorbance there is
below the threshold, and O2 now is off. At 535 nm, both FGc
and DTEc absorb. This results in switching O3 on and O0 off.
Output O1 is off because although FGc is emissive, the excited
state is quenched by the closed form of DTE as discussed above,
and emission intensity is far below the threshold. Thus, the
fourth line of the truth table is represented.
Thus, the decoder function of triad 1 has been demonstrated.
The tops of the black bars in Figure 7 show the signal-to-noise
level achieved in the measurements, which is far greater than
necessary to meet requirements. Absorbance and emission were
both monitored for 5 s. When I0 and I1 are applied simulta-
neously, the output amplitudes are a weak function of the order
of application of the inputs. The variation due to this factor is
shown in Figure 7 and is clearly too small to affect function.
Photocycling of the Decoder. Figure 8 shows the results of
experiments to determine the photostability of the triad under
decoder operating conditions. Each input-output combination
was preceded by a reset operation with green light. There is a
slight amount of performance degradation after five cycles. This
is due mostly to photodecomposition of DTE.
Conclusions
This research shows that a single molecular species can act
both as a 4-to-2 encoder and a 2-to-4 decoder, using only optical
inputs and outputs. In addition to performing a relatively
complex mathematical operation, the triad can be reconfigured
in situ simply by enabling a different set of inputs and outputs.
This ability to be reconfigured is often a characteristic of
chemical logic gates. The use of only optical inputs and outputs
means that no material access to the gate element is necessary,
permitting in principle monolithic three-dimensional arrays of
devices. Achieving the encoder and decoder functions in a single
molecule is logically complex. In the decoder, for example, only
one of the outputs is switched on for each input combination.
This means that I0 must counteract the effect of I1, and vice
versa. At the same time, applying both I0 and I1 must give rise
to a unique new output. This means that the molecule must
function simultaneously as one NOR gate, two complementary
INHIBIT gates, and one AND gate that all share inputs and
initial state. The complexity of function and the reconfigurability
are both due in part to the fact that unlike electronic devices,
where all inputs and outputs are electronic, molecular photonic
systems can use light of different wavelengths, allowing for
multiple inputs and outputs that can travel through the same
spatial regions.
One could imagine a variety of uses for photonic molecular
encoders, decoders, etc., in addition to the kinds of applications
now satisfied by their electronic analogs. As mentioned earlier,
the times necessary for the various photoisomerization reactions
depend on the intensity of the light used; the actual photochemi-
cal processes occur during the lifetimes of the photochrome
excited states (nanoseconds or less). Much faster switching than
that reported in Table 2 would be possible with more intense
irradiation. That said, it is unlikely that molecular systems such
as that described here will replace traditional electronics for
current uses in computing and data communication in the near
Figure 7. Performance of the triad as a 2-to-4 decoder. The black bars
show the signal amplitude at the appropriate monitoring wavelengths for
outputs O0 through O3 when the inputs are off, when I0 is on, when I1 is
on, and when both are on. The dashed lines are thresholds used to distinguish
an on output from an off output. The variation in height at the top of each
bar shows the variation in signal amplitude during measurement for 5 s.
The small bars to the right of the tops of the outputs when both I0 and I1
are on show the variation seen when the order of application of the two
inputs is reversed.
virtually no absorbance at wavelengths >470 nm, and with the
proper choice of a threshold level (Figure 7) the absorbance at
393 nm is low enough to give an off response for O2. Only low
emission intensity is observed at 624 nm upon excitation at 500
nm. Thus, the only output that is on is O0, as the transmittance
at 535 nm is close to unity. The system thus represents the first
line of the truth table for the decoder.
When I0 is turned on, the sample is isomerized mainly to
FGc-DTEo. As can be seen from Figure 4, the presence of
the visible absorption band of FGc (λmax ∼ 511 nm) leads to
some absorbance at 535 nm. The absorbance is still below the
threshold, so that O3 is still off, but strong enough to reduce
the transmittance at 535 below the threshold, so that O0 is now
off. The absorbance at 393 also decreases, leaving O2 off.
However, fluorescence emission from FGc is now observable
at 624 nm, and O1 is on. This corresponds to the second line of
the decoder truth table.
Alternatively, applying input I1 at 302 nm isomerizes the triad
mainly to FGo-DTEc. The absorbance at 393 nm rises above
the threshold because both chromophores absorb there (Figure
9
11126 J. AM. CHEM. SOC. VOL. 130, NO. 33, 2008

Molecular All-Photonic Encoder-Decoder
A R T I C L E S
Figure 9. Structures of synthetic intermediate fulgide 4 and fulgimide 5.
samples were degassed by six freeze-pump-thaw cycles to a final
pressure of approximately 10-5 Torr. The absorption measurements
were performed using a CARY 4000 UV-vis spectrometer. A
SPEX Fluorolog τ2 was used for the emission measurements. After
irradiation with the different light input combinations, the inputs
were turned off, and the absorbance and the emission were
monitored separately using the instruments described above. The
sample concentration was ∼2.5 × 10-5 M. The 397 nm light and
the broad-band green light were generated by a 1000 W Xe/Hg
lamp operated at 450 W and equipped with a hot mirror (A ) 1.8
at 900 nm) to reduce the IR intensity. A 397 nm interference filter
was used for the 397 nm light, and a VG 9 glass filter (A < 1.5
between 460 and 590 nm) was used for the broad-band green light.
The resulting light power densities were ∼1.6 mW/cm2 and ∼9
mW/cm2 for the 397 nm light and the green light, respectively.
The 302 and the 366 nm UV inputs were generated by UVP hand-
held UV lamps (models UVM-57 and UVGL-25, respectively), each
having power densities of 1.5 mW/cm2. The total sample volume
was ∼3 mL. With the use of the 397 nm light and the broad-band
green light, only one-third of the sample volume was exposed to
the light at any one time, whereas the whole sample volume was
exposed to 302 and 366 nm light. The samples were stirred
continuously during all irradiation processes.
Synthesis. The preparation of 1,3-bis(bromomethyl)-5-iodoben-
zene and 1-iodo-3,5-bis(N-phthalimidomethyl)benzene have been
previously described,27 as has that of fulgide 4 (Figure 9).28 1
H
NMR spectra were recorded on Varian Unity spectrometers at 300
or 500 MHz. Samples were dissolved in deuteriochloroform, with
tetramethylsilane as an internal reference, or in toluene-d8. Mass
spectra were obtained on a matrix-assisted laser desorption ioniza-
tion time-of-flight spectrometer (MALDI-TOF).
Figure 8. Photocycling of triad 1 in the decoder mode. The signal amplitude
for each of the four outputs is shown following five repetitions of the
sequence of inputs: reset (R), I0 (0), reset, I1 (1), reset, both I0 and I1 (B).
The sequence is indicated at the bottom of the figure for the first six
applications. The dashed lines are thresholds used to distinguish an on output
from an off output. There are small changes over the course of the
irradiations that indicate a small amount of decomposition of DTE.
Dyad 5. To a suspension of 113 mg (0.22 mmol) of 1-iodo-3,5-
bis(N-phthalimidomethyl)benzene in 15 mL of ethanol was added
0.7 mL of hydrazine hydrate, and the solution was refluxed under
nitrogen for 3 h. The solution was cooled to room temperature,
and 30 mL of 2 M HCl was added. After 15 min the solution was
poured into 150 mL of 1 M NaOH, and the resulting mixture
was extracted four times with dichloromethane. The organic layer
was washed with brine and dried over anhydrous K2CO3. The
solvent was evaporated, and to the residue was added 140 mg (0.49
mmol) of fulgide 4 and 15 mL of toluene. The solution was stirred
under nitrogen at 65 °C overnight. The next day 40 mg (0.29 mmol)
of anhydrous ZnCl2 and 400 µL (1.9 mmol) of hexamethyldisilazane
were added, and the mixture was again stirred overnight at 65 °C
under nitrogen. The reaction mixture was poured into water, and
the resulting mixture was extracted twice with toluene and once
with dichloromethane. The organic extracts were combined, and
the solvent was distilled at reduced pressure. The residue was
dissolved in dichloromethane and chromatographed on silica gel,
eluting with 25% ethyl acetate in hexanes, to yield 22 mg (0.026
future. However, molecular systems can function in applications
where traditional electronics cannot. Examples are the use of
optically interrogated molecular logic to label nanoparticles26
and in a “molecular keypad lock” that could have applications
in drug delivery.16 For the system discussed here, if a sample
of nano-objects labeled with suitable decoder molecules were
introduced into an organism or other complex system, and each
of two spatially distinct sites in the organism were irradiated
with one of the two inputs, one could later interrogate the nano-
objects with light and determine whether they had visited both,
either one, or neither of the irradiated locations simply by
reading the outputs. Similar applications on larger scales can
also be envisioned.
Experimental Section
Spectroscopic Measurements. Distilled 2-methyltetrahydrofuran
was used as the solvent for spectroscopic measurements. The
(27) Ryan, T. J.; Lecollinet, G.; Velasco, T.; Davis, A. P. Proc. Natl. Acad.
Sci. U.S.A. 2002, 99, 4863–4866.
(28) Liang, Y.; Dvornikov, A. S.; Rentzepis, P. M. J. Photochem.
Photobiol., A 1999, 125, 79–84.
(26) de Silva, A. P.; James, M. R.; McKinney, B. O. F.; Pears, D. A.;
Weir, S. M. Nat. Mater. 2006, 5, 787–790.
9
J. AM. CHEM. SOC. VOL. 130, NO. 33, 2008 11127

A R T I C L E S
Andre´asson et al.
mmol, 12%) of dyad 5. 1H NMR (300 MHz, CDCl3): δ 1.23 (6H,
m), 1.94 (3H, s), 1.98 (3H, s), 2.47 (6H, m), 3.89 (3H, s), 3.67
(3H, s), 4.75 (4H, m), 7.11-7.19 (3H, m), 7.25-7.32 (3H, m),
7.48 (2H, s), 7.54-7.62 (3H, m), 7.67 (2H, m). MALDI-TOF: calcd
for C44H41IN4O4 + Na, 839.2; obsd, 839.3. Isotopic distribution:
calcd m/z (relative intensity) 839.215 (59), 840.218 (33), 841.221
(8); found 839.316 (67), 840.246 (29), 841.209 (4).
Triad 1. A mixture of 2 mL of tetrahydrofuran and 2 mL of
triethylamine was bubbled with argon for 1 h, after which time 22
mg (0.026 mmol) of 5 was added along with 30 mg (0.052 mmol)
of 1-[5-(4-methoxyphenyl)-2-methylthien-3-yl]-2-[2-methyl-5-(4-
ethynylphenyl)thien-3-yl]-3,3,4,4,5,5-hexafluorocyclopentene (3),29
8 mg (0.026 mmol) of Ph3As, and 2 (0.002 mmol) mg of Pd2(dba)3.
The reaction vessel was capped with a septum, and the mixture
was stirred overnight at room temperature. The solvent was then
evaporated, and the residue was dissolved in dichloromethane and
subjected to preparative TLC (2:3 ethyl acetate/hexanes). The
appropriate band was scraped off, and the product was extracted
from the silica gel with acetone and dichloromethane. The solvent
was evaporated, and the residue was again purified by preparative
TLC, this time eluting with 1% ethyl acetate in dichloromethane
to give 21 mg (0.017 mmol, 64%) of triad 1. 1H NMR (300 MHz,
CDCl3): δ 1.25 (6H, m), 1.96 (12H, m), 2.47 (6H, m), 3.44 (3H,
s), 3.68 (3H, s), 3.84 (3H, s), 4.83 (4H, m), 6.92 (2H, d, J ) 9
Hz), 7.10-7.21 (4H, m), 7.25-7.33 (4H, m), 7.37 (1H, br s), 7.48
(2H, d, J ) 9 Hz), 7.52 (6H, m), 7.54 (1H, s), 7.56-7.64 (3H, m).
MALDI-TOF: calcd for C74H60F6N4O5S2 + Na, 1285.4; obsd,
1285.4. Isotopic distribution: calcd m/z (relative intensity) 1285.380
(40), 1286.383 (34), 1287.384 (19), 1288.385 (8); found 1285.377
(45), 1286.342 (35), 1287.308 (15), 1288.331 (6).
Acknowledgment. This work was supported by the National
Science Foundation (CHE-0352599) and the Swedish Research
Council.
(29) Liddell, P. A.; Kodis, G.; Moore, A. L.; Moore, T. A.; Gust, D. J. Am.
Chem. Soc. 2002, 124, 7668–7669.
JA802845Z
9
11128 J. AM. CHEM. SOC. VOL. 130, NO. 33, 2008