A simple and robust PET-based anthracene-appended O-N-O chelate for sequential recognition of Fe3+/CN– ions in aqueous media and its multimodal applications

Article abstract of DOI:10.1016/j.ica.2018.07.007

A very simple, robust, and pico/nanomolar-sensitive 9,10-diethanolamine-substituted fluorescent Fe³⁺/CN^– probe (PD) was synthesised, and its sensing abilities towards various ions were studied in mixed aqueous media. PD selectively recognised Fe³⁺ ions through a ‘turn on’ response with an excellent binding constant (K_a, 9.29 × 10⁶ M^?1) in 1:2 binding stoichiometry at pH 7.0 in phosphate-buffered saline (PBS). The in situ generated Fe³⁺·PD ensemble sequentially recognised CN^– ions with an excellent binding constant (K_a 1.72 × 10⁸ M^?1) via a ‘turn off’ mode by extruding Fe³⁺ ions from the ensemble. The highly selective sequential ‘on-off’ responses towards Fe³⁺ and CN^– ions were attributed to inhibition and restoration of photoinduced electron transfer (PET) and chelation-induced enhanced fluorescence (CHEF) effects from the chelating N and O heteroatoms. PD was able to detect Fe³⁺ and CN^– ions in real water samples satisfactorily at picomolar to sub-nanomolar levels. A colorimetric assay based on pyrocatechol violet (PCV) was also able to detect Fe³⁺/CN^– in a sequential manner (up to sub-micromolar level) by a change in colour from colourless to yellow/pale green without any interferences from other ions. Based on the complexation and decomplexation mechanism, bio-imaging photonic INHIBIT logic circuit strips were prepared for use under physiological conditions. In addition, solid-phase recognition of Fe³⁺/CN^– ions was demonstrated using cost-effective paper-based strips.

Full text of DOI:10.1016/j.ica.2018.07.007

Accepted Manuscript
Research paper
A Simple and Robust PET-Based Anthracene-Appended O-N-O Chelate for
Sequential Recognition of Fe3+/CN- Ions in Aqueous Media and Its Multimodal
Applications
Anup Pandith, Jun-Hyeak Choi, Ok-Sang Jung, Hong-Seok Kim
PII:
S0020-1693(18)30642-X
https://doi.org/10.1016/j.ica.2018.07.007
ICA 18353
DOI:
Reference:
Inorganica Chimica Acta
To appear in:
Received Date:
Revised Date:
Accepted Date:
11 May 2018
3 July 2018
3 July 2018
Please cite this article as: A. Pandith, J-H. Choi, O-S. Jung, H-S. Kim, A Simple and Robust PET-Based Anthracene-
Appended O-N-O Chelate for Sequential Recognition of Fe3+/CN- Ions in Aqueous Media and Its Multimodal
Applications, Inorganica Chimica Acta (2018), doi: https://doi.org/10.1016/j.ica.2018.07.007
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

A Simple and Robust PET-Based Anthracene-Appended O-N-O Chelate for
Sequential Recognition of Fe3+/CN- Ions in Aqueous Media and Its Multimodal
Applications
Anup Panditha, Jun-Hyeak Choia, Ok-Sang Jungb, and Hong-Seok Kima*
aDepartment of Applied Chemistry, School of Applied Chemical Engineering,
Kyungpook National University, Daegu 41566, Republic of Korea
bDepartment of Chemistry, Pusan National University, Busan 46241, Republic of Korea
Corresponding author: Tel: +82 53 9505588; Fax: +82 53 9506594
E-mail address: kimhs@knu.ac.kr
Abstract: A very simple, robust, and pico/nanomolar-sensitive 9,10-diethanolamine-substituted
fluorescent Fe3+/CN- probe (PD) was synthesised, and its sensing abilities towards various ions were
studied in aqueous media. PD selectively recognised Fe3+ ions through a ‘turn on’ response with an
excellent binding constant (Ka, 9.29  106 M-1) in 1:2 binding stoichiometry at pH 7.0 in phosphate-
buffered saline (PBS). The in situ generated Fe3+PD ensemble sequentially recognised CN- ions with an
excellent binding constant (Ka 1.72  108 M-1) via a ‘turn off’ mode by extruding Fe3+ ions from the
ensemble. The highly selective sequential ‘on-off’ responses towards Fe3+ and CN- ions were attributed
to inhibition and restoration of photoinduced electron transfer (PET) and chelation-induced enhanced
fluorescence (CHEF) effects from the chelating N and O heteroatoms. PD was able to detect Fe3+ and CN-
ions in real water samples satisfactorily at picomolar to sub-nanomolar levels. A colorimetric assay
based on pyrocatechol violet (PCV) was also able to detect Fe3+/CN- in a sequential manner (up to sub-
micromolar level) by a change in colour from colourless to yellow/pale green without any interferences
from other ions. Based on the complexation and decomplexation mechanism, bio-imaging photonic
INHIBIT logic circuit strips were prepared for use under physiological conditions. In addition, solid-phase
recognition of Fe3+/CN- ions was demonstrated using cost-effective paper-based strips.
1

Keywords: Photoinduced electron transfer, Chelation-induced enhanced fluorescence, Bio-imaging and
Photonic INHIBIT logic circuit.
1. Introduction
Ferric (Fe3+) ions are one of the most abundant metal ions and play a vital role in biological systems,
such as in oxygen transport, as well as various metabolic processes and their regulation [1,2]. Most
reported fluorescent/colorimetric probes for sensing Fe3+ ions operate in organic or mixed
organic/aqueous media (including real samples analysis) with HEPES/Tris-HCl buffers [3-4]. However, in
majority of the cases, fluorogenic-image-based systems require cell culture under neutral conditions,
and generally, contain phosphate buffers because of their better biocompatibility relative to other
buffers; that is, phosphate buffers function in the neutral pH range and do not disrupt normal cellular
biochemical processes. Phosphates (inorganic/organic) are major ionic species in the biological fluids of
pro/eukaryotic organisms and play a vital role in metabolic process such as ion/electron transport,
osmotic regulation, energy storage, and energy utilization [5]. In contrast, the intracellular tracking of
biologically important Fe3+ and CN- ions through complexation and decomplexation processes using
photonic switches is challenging because of the high probability of interference from the numerous
components of biological fluids. Accordingly, various research groups have reported highly efficient
chromo-fluorogenic probes for Fe3+ ions in cellular/subcellular systems [6]. Such highly selective probes
are typically very large molecules requiring complicated synthetic routes (multi-step), polymers [7-9], or
graphene/carbon-based quantum dots [10-12], most of which are ‘turn off’-based sensors, functioning
through the general optical inner filter effect (OIE). In contrast, organic/inorganic nanoparticle hybrids
and metal-organic frameworks (MOFs) [13-15], supramolecular assemblies (calixarenes, rotaxanes, etc.)
[16], and covalent organic frameworks (COFs) [17] are selective according to the size of the cages and
ionic radius of the frameworks and the metal ions. However, their synthesis is complex and time-
2

consuming, and they are expensive to characterise and have other limitations. In addition, few research
groups have reported the effects or limitations of the buffer on the detection selectivity and sensitivity
for ionic analytes under physiological conditions [18,19]. On the other hand, most ‘turn on’ fluorescent
probes are based on small molecules and suffer from interference via electron-transfer-aided quenching
(dn) from Cu2+, Ni2+, and Co2+ ions because of their similar sizes and chelation capabilities to those of
Fe3+ ions, in addition to hard metal (Al3+) interferences [20-22]. Recently, ethylene-dicarboxylic-acid-
appended anthracene-based Cu2+ and Fe3+ chelates have been used to detect the anions through a
‘switch on’ response in phosphate buffer [23]. These studies have revealed that Fe3+ sensing must be
carried out at an acidic pH (pH 3.5) to keep the carboxylic acid active during the chelation process.
However, the observed detection limits, analysis conditions, and ‘switch off’ response limit its
applications towards cellular/subcellular tracking studies. In contrast, a diethanolamine unit has been
utilised to design various fluorescent probes via the incorporation of various functional groups, as well
as a heterocyclic unit with chromophore/fluorophore molecules. Based on our current research
interests and to investigate the structural-change-aided photophysical properties, we have designed
various small fluorescent probes based on polyaromatic hydrocarbon skeletons, such as anthracene,
pyrene, and perylene, for the detection of neutral molecules and ions in organic or aqueous media [24-
26]. However, the obtained results do not explain the sensing ability of the O-N-O unit of
diethanolamine towards various metal and anions in the biologically more useful phosphate buffered
saline (PBS) as compared to HEPES buffer. Continuing our work, we investigated the behaviour of the X-
N-X (X = Cl or Br) unit towards chelation using aminomethylanthracene, a well-known optical signalling
unit, in saline buffered media. Such studies could give insights into the photoinduced electron transfer
(PET)-based sensing of the aminomethylanthracene unit through the spatially coupled electron-rich
orbitals of heteroatoms with hydrogen bonding capabilities. Furthermore, the solvent/metal-assisted
photophysical effects, as well as efficiency, can be used to compare structurally similar reported
3

receptor units. This would allow for the strategic design of highly efficient and novel small chromo-
fluorogenic probes for specific anionic analytes under physiological conditions, as well as in vivo studies.
Because of the high toxicity of CN- ions, which results from the rapid inhibition of cytochrome oxidase
and subsequent cellular anoxia, its sensing and cellular tracking is essential for safety [27]. In general, a
non-toxic-metal-based displacement approach is useful and advantageous because of its ease of
preparation and handling, as well as the low probability of cell death during endocytosis-triggered
analysis. Accordingly, in this paper, we report a highly selective ‘off-on-of’ sequential photonic switch
that is chemically driven by the addition of Fe3+/CN- ions, which we studied in aqueous media (PBS), as
well as the effective demonstration of intracellular tracking in primitive eukaryotic cell lines. We also
briefly demonstrate the use of a pyrocatechol-violet (PCV)-based colorimetric assay, real sample analysis,
and cost-effective paper-strip-based sensing with the aid of a UV lamp. The sensing abilities of the
designed probes are demonstrated through UV-Vis, fluorescence, and 1H-NMR studies and supported by
density functional theory calculations.
2. Experimental
2.1. Materials and methods
1H and 13C NMR spectra were recorded on a Bruker AM-500 spectrometer. High-resolution fast-atom
bombardment mass spectrometry (FAB-HR) was carried out at the Korea Basic Science Institute (KBSI)
Daegu Branch. UV-Vis absorption spectra were recorded on a Shimadzu UV-1650PC spectrophotometer.
Fluorescence spectra were measured using 1-cm quartz cells with an excitation and emission slit width
of 3/5 on a Shimadzu RF-5301 fluorescence spectrometer equipped with a xenon discharge lamp. FT-IR
spectra were recorded on a Perkin-Elmer spectrometer using the attenuated total reflection (ATR)
method. All measurements were performed at 25 °C. Analytical-grade acetonitrile (ACN) was purchased
4

from Merck. Diethanolamine, PCV, and other compounds used throughout the synthesis were
purchased from Sigma–Aldrich Co. and were used as received. 9,10-Bis(chloromethyl)anthracene was
prepared according to a literature procedure [28]. PCV was initially dissolved in ACN and diluted with
specific water fractions (at appropriate pH values) for analysis. Fluorescence quantum yields were
determined by integrating the corrected fluorescence spectra. Quinine hemisulfate (0.5 M, H2SO4) was
used as the fluorescent standard to correct all fluorescence spectra (φ = 0.54). The fluorescence spectra
were recorded immediately.
2.2. Association/Stern–Volmer quenching constants and limit of detection calculations
The association constants for the binding of PD to various targets were calculated from the non-linear
curve fitting of fluorescence titration data using Gnuplot (Version 4.0) and the reduced -squared
method (error bound within ±10.0%) adapted from the adjusted r-squared method [29]. The Stern–
Volmer quenching constants, Ksv, with CN- and PCV were calculated from fluorescence titration data
using the equation, I0/I = 1 + Ksv [Q] (for linear fitting) and [I0/I] = e(a+KsvX+b) (for non-linear fitting), where I0
and I are the fluorescence intensities of the probe in the absence and presence of the respective
analytes, respectively; [X] is the analyte concentration; a is a constant; and b is a function of X based on
polynomial fitting (in second-order exponential values because of the consideration of the average of
various linear subdivided intervals) and considered constant. Solutions (1.0/0.1 × 10-6 M) of the
ensemble PDFe3+ (1:2) were prepared in 50-mL volumetric flasks (±0.025 mL). The limit of detection
(LOD) was calculated based on the results determined by the fluorescence titration studies, where
observable changes occurred following the addition of the analytes. The LOD was calculated using the
standard equation 3/m, where  is the standard deviation, and m is the slope of the curve of the linear
relationship between the fluorescence intensity and the analyte concentration with a signal-to-noise
(S/N) ratio of 5. The limit of detection was calculated after calibrations were performed.
5

2.3. Synthesis of probes
Synthesis of PD: To a stirred solution of 9,10-bis(chloromethyl)anthracene (0.1 g, 0.36 mmol) in ACN
(10 mL), diethanolamine (0.19 g, 1.80 mmol), anhydrous K2CO3 (0.13 g, 0.91 mmol), and KI (0.15 g, 0.90
mmol) were added under N2. The reaction mixture was refluxed for 20 h. Then, the reaction mixture was
filtered carefully and concentrated to dryness, washed with water, and extracted with EtOAc. The crude
product was purified using silica-gel chromatography (elution with 15% MeOH in CH2Cl2:EtOAc (7:3
v/v%) containing 2.5% NH4OH, Rf = 0.45), yielding a yellow solid (0.13 g, 86%). 1H NMR (500 MHz, DMSO-
d6) δ 2.65 (t, J = 6.5 Hz, 8H, Hb), 3.43 (q, J = 9.0 and 3.0 Hz, 8H, Ha), 4.30 (t, J = 3.0 Hz, 4H, -OH), 4.61 (s,
4H, Hc), 7.52-7.54 (m, 4H, He-Ant), 8.61-8.63 (m, 4H, Hd-Ant). 13C NMR (125 MHz, DMSO-d6) δ 51.4, 56.0,
59.3, 125.1, 125.6, 125.8, 130.6, 130.9, 130.5. Electrospray ionization (ESI)-HR mass spectrometry: m/z
calculated for [M]+ C28H32N6, 412.2361; found, 412.2361.
Synthesis of PM: 9-Bromomethylanthracene was prepared in good yield (>87%) by the reduction of 9-
anthracenecarboxaldehyde, followed by bromination with PBr3 under N2 atmosphere in toluene. To a
stirred solution of 9-bromomethylanthracene (0.1 g, 0.37 mmol) in ACN (10 mL), diethanolamine (0.07 g,
0.67 mmol), anhydrous K2CO3 (0.07 g, 0.50 mmol), and KI (0.08 g, 0.50 mmol) were added under N2. The
reaction mixture was refluxed for 12 h. The work-up was similar to that of PD. The crude product was
purified using silica-gel chromatography (elution with 8.0% MeOH in CH2Cl2:EtOAc (7:3 v/v%) containing
1.4% NH4OH, Rf = 0.5), yielding a brownish semi-solid PM (0.10 g, 91%). 1H NMR (500 MHz, DMSO-d6) δ
2.65 (t, J = 7.5 Hz, 4H), 3.44 (broad, 4H), 4.30 (bs, -OH, 2H), 4.62 (s, AnCH2, 2H), 7.49–7.57 (m, 4H), 8.07–
8.09 (m, 2H), 8.58 (m, 3H). 13C NMR (125 MHz, DMSO-d6) δ 51.2, 56.1, 59.3, 125.0, 125.2, 125.7, 127.1,
128.8, 130.6, 130.9, 131.0. FAB-HR mass spectrometry: m/z calculated for [M+H]+ C19H21NO2, 296.1651;
found, 296.1654.
6

Synthesis of PM-Cl: To a stirred solution of PM (0.25 g, 0.85 mmol) in toluene, thionyl chloride (0.5g,
4.24 mmol) was added slowly under N2 at 0 °C. The reaction mixture was further stirred for 3 h, and the
temperature was gradually increased to room temperature. Then, the reaction mixture was evaporated
to dryness and quenched with H2O extracted with EtOAc. The obtained EtOAc layer was washed with
brine, followed by drying over Na2SO4. The obtained crude product was crystallised in toluene, yielding a
brownish yellow solid (0.12 g, 65%). 1H NMR (500 MHz, DMSO-d6)  3.04 (t, J = 5.0 Hz, 4H), 3.43 (t, J = 6.5
Hz, 4H), 7.39–7.50 (m, 4H), 7.93–7.95 (m, 3H), 8.31–8.37 (m, 4H). 13C NMR (125 MHz, DMSO-d6) δ 38.3,
51.3, 56.1, 124.4, 125.1, 125.7, 127.2, 129.2, 131.3, 131.4, 133.5, 134.1; FAB-HR mass spectrometry: m/z
calculated for [M+H]+ C19H19NCl2, 331.0896; found, 331.0895.
2.4. Preparation of probe and ensemble-coated paper strips
Paper strips (5  2 cm) were coated with PD/PDFe3+ (1:2, 5  10-6 M, CH3CN-H2O (3:1, v/v)), followed
by the removal of the solvent under vacuum at room temperature. An ensemble of coated filter papers
was then cut into pieces (1  1 cm) to produce test strips for the detection of Fe3+/CN- ions. All images
were processed using Image J [30].
2.5. Theoretical calculations
Density functional theory (DFT) calculations were performed according to previously reported methods
[31].
2.6. Biological studies
2.6.1. Intracellular imaging
To demonstrate the feasibility of PD and the PDFe3+ (1:2) ensemble for sensing Fe3+/CN- in complex
matrices, an intracellular detection study was performed. Thus, different concentrations of Fe3+/CN- ions
were introduced to PD/PDFe3+ (1:2) ensemble-incubated KCTC-11282 cells (KCTC: Korean cell type
7

culture collection) to obtain the appropriate optimised analyte concentration. After 45 or 120 min of
incubation, the cells were washed with a buffer solution to remove the free PD/PDFe3+ (1:2)/Fe3+/CN-
and any nonspecific adsorption on the cell surface. Subsequently, confocal fluorescence microscopic
images of the KCTC-11282 cells were recorded.
2.6.2. Cell viability
To determine the cell cytocompatibility of the prepared PD/PD●Fe3+ (1:2), a standard methyl thiazolyl
diphenyl tetrazolium bromide (MTT) assay and live-cell imaging were performed using KCTC-11282 cell
lines. The KCTC-11282 cells were cultured in a yeast malt extract medium. After 24 h, the cells were
incubated with different concentrations of PD/PDFe3+ (1:2) (0–100.0 M), and then cultured for
another 24 and 48 h. Each (single) concentration added to three wells was taken as a parallel control,
and wells containing cells that had not been treated with PD/PDFe3+ (1:2) were used as negative
controls. Some wells were also measured without cells and treated as blank controls. After an additional
24 or 48 h incubation period, MTT (20.0 μL) was added to each well, and the optical density was
measured at 570 nm. Cell viability was calculated using the following formula:
Cell viability (%) = (Asample − Ablank) × 100%/(Acontrol − Ablank).
Here, ‘Asample’ is the absorbance of a well with cells, MTT solution, and PD/PDFe3+; ‘Ablank’ is the
absorbance of a well with medium and MTT solution without cells, and ‘Acontrol’ is the absorbance of a
well with MTT solution and cells without PD/PD●Fe3+ (1:2). The data are expressed as the percentages
of viable cells compared to the survival of the control (untreated cells taken as controls showing 100%
viability). Three parallel wells were run for each concentration, and each experiment was repeated three
times. Similarly, measurements involving PD/PDFe3+ (1:2) in the presence of CN- were performed.
3. Results and discussion
8

In general, because of the formation of highly stable M3+–phosphate complexes (very low solubility
product) in aqueous media, phosphate buffers are unsuitable for the fluorescence imaging of hard metal
ions. In general, fluorescent probes designed for hard metal recognition can be used under in vitro
conditions; however, they often fail to recognise hard metals in cellular systems in phosphate buffer [32-
35]. The diethanolamine unit is one of the simplest chelating units, and the O-N-O group separation
through ethylene spacers can yield stable chelates with transition metals depending on the nature of
the solvent and metal hardness (positive charge density) because of various flexible C-C rotation across
the ethylene spacers. The 9,10-disubstituted anthracene analogues show extensive variation in their
photophysical properties in the excited state, typically induced by the solvents and other
physicochemical parameters such as temperature, pressure, salt conditions, and the nature of the
substituent groups. Such factors generally affect the vibronically coupled anthracene monomer bands
via PET, twisted intramolecular charge/electron transfer (TI-C/E-T), energy/electron transfer (ET),
resonance energy transfer (RET), and chelation-induced enhanced fluorescence (CHEF) [36,37]. In
general, the PET-based system shows profound quenching in polar protic media, mainly arising from the
easier delocalization of N (through  bond) and O (through space) heteroatoms, and this could be used
to prepare novel ‘turn on’ probes for specific metal ions (inhibiting PET along with CHEF effects) and
‘turn off’ responses with anions (pronounced PET efficiency through weakening of the O-H bond
strength) achieved through solvent-aided metal binding and extrusion, respectively. On the other hand,
the symmetric incorporation of hydrophilic/lipophilic-spacer chelating units conjugated with a
hydrophobic fluorophore can balance the lipophilic-hydrophobic-hydrophilic characteristics of the probe,
which could enhance endocytosis through the cell membrane without damage or hindrance. To evaluate
the significance of hydrophilicity and chelation sensitivity, O-N-O-containing, mono-armed probes PM
and PM-Cl were synthesised. These studies also revealed the through-bond and through-space electron
transfer effects on the excited state photophysical behaviour of 9,10-disubstituted anthracene
9

analogues in mixed aqueous media, opening new possibilities for next-generation anthracene-based
sensors.
Fig. 1. Structural diversity of the PD probe.
The PD probe was synthesised in good yield from commercially available anthracene via
chloromethylation, followed by the reaction with diethanolamine under a N2 atmosphere in the
1
presence of K2CO3 and KI. The structure of PD was confirmed by H and 13C-NMR, single-crystal X-ray
1
diffraction analysis (XRD), and FAB-HR mass spectroscopy (see ESI). The H NMR of PD in DMSO-d6
showed an AnCH2N(CH2CH2OH)2 peak at  4.60 and -OH protons, as a triplet, at  4.29. In the aromatic
region, only two sets of protons (four protons each) at  7.51-7.53 and 8.60–8.62 ppm as a multiplet
were observed. The 13C NMR spectrum contained peaks at  51.4 for AnCH2NCH2 and  59.3 ppm for
CH2OH, along with four peaks in the aromatic region, which indicated the purity of the compound. In
addition, electron ionization (EI+)-HR mass spectra showed a peak with m/z 412.2361 for [M]+. Similarly,
the mono-armed probes were synthesised from commercially available 9-anthracenecarboxaldehyde via
NaBH4 reduction, followed by SN2 coupling with diethanolamine in the presence of K2CO3 and KI, yielding
the PM probe, which was subsequently chlorinated with SOCl2 under a N2 atmosphere, yielding PM-Cl
(see Supplementary Information for details). In contrast, the stability of the bromo derivative was low
10

for photophysical studies in mixed aqueous media, and thus, this probe was not studied in detail. All the
described spectral features clearly confirm the successful formation of the probes.
Scheme 1. Reaction conditions: (i) H2NCH2CH2OH, K2CO3, KI, CH3CN, reflux 32 h, (ii) NaBH4, EtOH, rt, 3 h,
(iii) PBr3, toluene, KHCO3, 0 °C to rt, 5 h (iv) H2NCH2CH2OH, K2CO3, KI, CH3CN, reflux 19 h, and (v) SOCl2,
CH2Cl2, rt, 16 h.
The initial sensing ability of the PD probe was tested in polar aprotic ACN because of solubility issues,
as well as electron transfer effects. The UV-Vis spectra of the probe (1.0 M) in ACN showed the typical
anthracene absorption bands at 323, 339, 356, 375, and 395 nm (Fig. S1). The absorption bands 339, 356,
375, and 395 nm (log  4.550, 4.578, 5.100, and 5.107, respectively) were assigned to the ‘spin allowed
and symmetry forbidden’ A1gB2U (S0S1) -* and n-* transitions. In contrast, the high-energy
absorption band at 323 nm resulted from S0S2 transitions. Upon the addition of 10.0 eq. of various
metal ions such as Zn2+, Cu2+, Pb2+, Ag+, Ni2+, Cd2+, Co2+, Hg2+, Fe2+, Fe3+, Ga3+, Cs+, In3+, Al3+, Mg2+, Ca2+, Na+,
Rb+, and K+, only Fe3+ resulted in a red shift with increased molar absorptivity at the respective
absorption maxima with ±2 to ±3 nm proximal shifts. The molar absorptivities (log ) of the PD and
11

PDFe3+ complex changed from 4.230 to 4.920, 4.550 to 4.968, 4.878 to 5.102, 5.100 to 5.214, and 5.107
to 5.195 at 323, 339, 356, 375, and 395 nm, respectively. The substantial changes after the addition of
Fe3+ solutions suggest that significant interactions occurred between PD and Fe3+ in ACN (Fig. S2). Upon
excitation of a 0.1 M ACN solution of PD at 376 nm, three well-separated, vibronically coupled
emission bands at 406, 429, and 456 nm (exact mirror images of the absorption spectrum) appeared,
confirming that the obtained spectrum results from the anthracene monomer. Considering that probe
sensing in aqueous media is crucial, the probe was tested in water. As the water percentage of the 0.1
M solution of PD in ACN was increased slowly, a monotonic increase in emission intensity was
observed up to 75%, subsequently decreasing without any change in the spectrum shape at 406, 429,
454, and 485 nm (Fig. S3). The obtained results indicate the variations in the excited-state photostability
of the anthracene unit arising from hydrogen bonding after hydration. The significant increases in the
emission intensity after the addition of water result from the effective inhibition of PET (through-bond
and through-space) from heteroatoms such as N and O, upon hydrogen bonding. This process not only
hinders the electron transfer to highest occupied molecular orbital (HOMO) of photoexcited anthracene
but also inhibits C-C free rotation at the 9- and 10-positions in a symmetric manner, resulting in a
relatively rigid geometry in the presence of polar protic solvents with high proton donating capabilities.
The combined synergistic inhibition of PET and C-C rotations in a symmetrical manner is responsible for
the greater fluorescence in aqueous media.
To compare the PET efficiency towards the anthracene unit and chelating strength towards Fe3+ ions,
various derivatives (i.e., PM and PM-Cl) were synthesised, and their sensing abilities in ACN/ACN-H2O
media were analysed. The UV-Vis spectra of 1.0 M solutions of PM and PM-Cl showed characteristic
peaks similar to those of PD, (±3 nm range); upon excitation at 376 nm, peaks similar to that of PD (±5
nm) were obtained, but the emission intensities were different. The order of emission intensities, PM (
12

= 0.00121) > PM-Cl ( = 0.00234), clearly demonstrates that along with the methylene spacer on the N
atom, through-space PET from halogens is also significant in polar aprotic acetonitrile (Figs. S4 and S5).
The water tolerance of the sensitivity/photostability was also found to be higher for PD than PM, and
this could be seen in the UV-Vis and fluorescence spectra (Fig. S6). That is, the mono-armed halogenated
probe’s emission intensity increase at em = 426 nm was not significant, and the intensity reverted to the
original level at more than ~60% water, indicating that the photostability of PD was relatively higher
than that of the other probes because of the presence of -OH groups. Upon the addition of water, the
formation of hydrogen bonds with the O and N (via lone pair electrons) atoms also restricted the
rotation of the C-C bond, yielding structurally more rigid complexes and a gradual increase in the
fluorescence (Fig. 2).
Fig. 2. Schematic representation of the effect of water on PET inhibition and C-C bond rotation. Because
of the similar interactions in PD and PM, for simplicity, only a single-armed probe (PM) is shown.
Along with PET and CHEF, -linked heteroatoms and solvation through the multiple hydrogen bonds
and other non-covalent interactions can significantly affect the sensitivity and selectivity towards
specific metal ions, as well as the photophysical parameters. ACN is a polar aprotic, low-viscosity solvent
that can form strong hydrogen bonds with protons of the hydroxyl groups of the probe and enhance the
PET probability, resulting in very low emission. This process greatly increases the sensitivity of ‘turn on’
probes. Upon the addition of various solvents, such as dimethylsulfoxide (DMSO), dimethylformamide
13

(DMF), and 1,4-dioxane, comparatively lower emissions with negligible differences in absorption
compared to the polar protic solvents, such as MeOH, EtOH, and glycerol (GY) (Fig. S7), were obtained.
These results also confirmed the hydrogen donor effect towards PET. The solid-state crystal packing of
molecularly self-assembled PD (Fig. 3) shows the non-covalent interactions, such as multiple hydrogen
bonds, (3.10 Å) interactions, and OH-(2.90 Å)
CH-2.51 Å) interactions with relatively
shorter separations, further validating the feasibility of the aggregation-aided photoinduced electron
transfer process (Figs. S8 and S9, Table S1, for details CIF file see Supplementary Information) with very
low emission in the solid state.
Fig. 3. Self-assembled 2D and 3D molecular arrays of the PD probe obtained from single crystal analysis.
Details of the structure solution, as well as crystallographic information files (CIF), are incorporated in
the Supplementary Information (Tables S1 and S2).
Considering the practical use and the excited-state properties of the anthracene unit, PBS buffer (10X
diluted to 0.1 mM) was used for fluorescence studies. Because of the negligible difference in the UV-Vis
spectra of buffered media and ACN, fluorescence studies were carried out. Upon the addition of 10.0 eq.
of various metal ions (such as Fe3+, Al3+, Ga3+, In3+, Ca2+, Mg2+, Cu2+, Zn2+, Ni2+, Co2+, Fe2+, K+, Na+, Pb2+, Hg2+,
14

Ag+, Cd2+, Rb+, and Cs+) to the 1 mM solution of PD in ACN:H2O (3:7, v/v%) at pH 7.0 and at ex 376 nm,
only Fe3+, Fe2+ and Zn2+ showed responses: ~10-, 4-, and 2-fold enhancements in the emissions centred
at 406, 421, and 456 nm, respectively (Fig. 4). The mono-armed probes PM and PM-Cl showed a very
slight response and no response, respectively, in buffered ACN, indicating the selective chelation of Fe3+
ions (Fig. S10). Using non-linear regression analysis of the fluorescence titration data (Fig. 5) at em = 429
nm with Fe3+ ions, a Ka of 9.29  106 M-1 with an error of ±3.56% was obtained. Using a Job’s plot, we
confirmed the 1:2 binding stoichiometry between PD and Fe3+ ions from the maximum complex-
concentration-aided intensity at an Fe3+ mole fraction of ~0.67 (Fig. S11). The lowest detection limit for
Fe3+ ions was found to be 0.1 pM (Fig. S12). To validate the sensitivity, the mono-armed PM probe was
analysed under similar conditions, but the sensitivity was almost 100-times lower than that of PD in 80%
buffered medium (Fig. S13), demonstrating the complexation-strength-aided photostability of the 9,10-
disubstituted-anthracene analogues. In contrast, the halogenated compound PM-Cl did not show any
response towards Fe3+ ions, confirming the chelation strength and selectivity in aqueous media
attributed to hard acid–hard base interactions.
15

Fig. 4. Fluorescence studies of the PD probe (0.1 M) with 10.0 eq. of various metal ions in ACN:H2O (3:7,
v/v%) at pH 7.0 (PBS buffer).
Fig. 5. Fluorescence titration studies of PD (0.1 M) with Fe3+ ions (0.0 to 5.0 eq.) in ACN:H2O (3:7, v/v%)
at pH 7.0 at ex = 376 nm. The inset shows the fluorescence intensity at em = 429 nm vs. eq. of Fe3+ ions.
16

To evaluate the significance of binding capabilities and their strengths towards Fe3+ chelation, DFT
studies were carried out using the 6-31G* basis set. ACN and H2O were modelled as explicit ligands, and
H2O was used as an implicit solvent medium (Fig. 6). The calculated coordination bond lengths between
the Fe3+ ions and the O atoms of the hydroxyl groups labelled (i) and (iii) were 2.21 and 2.39 Å,
respectively. In contrast, the dative bond linkage between Fe3+ and the tertiary amine nitrogen atom
was measured to be (ii) 2.15 Å. Furthermore, the solvent molecules (one H2O and two ACN molecules)
stabilised the complex through the additional coordination sites, mainly through oxygen (iv) 2.09 Å and
nitrogen (v = 2.23 Å and vi = 2.29 Å) atoms. The interactions between the two Fe3+ ions with two
diethanolamine units at the 9- and 10-positions were found to be symmetric. The obtained band gap
energies (E = EHOMO – ELUMO) of the PD probe and PDFe3+ complex (1:2) were 3.39 and 3.22 eV,
respectively, and there was a substantial change in the HOMO and lowest occupied molecular orbital
(LUMO) towards more negative values, indicating that the obtained complex was thermodynamically
more stable with a lower electron density on the chelating atoms (Fig. 6b). Molecular electrostatic
potential maps (MEPs) obtained in H2O as an implicit and explicit solvent model further validated the
complexation hypothesis.
17

Fig. 6. a) DFT minimum energy structure of complex PD●(Fe3+)2, b) frontier molecular orbitals (FMO)
depictions showing the HOMO and LUMO, as well as band gap energies (E) in electronvolts, and c)
MEPs of the complex calculated using the 6-31G* basis set with ACN and H2O explicit ligands and H2O as
the implicit solvent medium (SM8 model).
FT-IR spectra were obtained using the ATR method and showed lower-intensity, shifted peaks at 3473
(O-H stretching), 2952 (C-H stretching), 1739 (C-H bending), 1534 (C=C), 1488 (C=C), 1465 (C=C), 1130 (C-
O stretching, split because of the merging with C-N stretching), and 780 cm-1 for PD, and additional
bands at 3284, 2950, 2887, 2881, 1655, 1525, 1446, 1351, 1018, and 660 cm-1 in PDFe3+ (1:2), indicating
the successful formation of the complex (Fig. S14). The FAB-HR mass spectrum contained a peak at m/z
666.9891, confirming the complexation process (Fig. S15). Fluorescence titration studies with Fe2+ ions
were also performed at pH 6.8 (Sorensen’s phosphate buffer, 0.1 mM) in the presence of 0.4 M SnCl2
18

to prevent auto-oxidation of Fe2+Fe3+ ions. The obtained binding constant Ka = 4.8  104 M-1 was lower
than that for the Fe3+ ions (Fig. S16), suggesting that complexation with Fe3+ was more stable and there
was no transfer of electrons from the metastable d5 orbitals, resulting in the inhibition of PET and
enhancement of CHEF effects. This also indicates the presence of a highly specific pocket for this
particular ion with specific solvent molecules in its coordination sphere. In contrast, the emission
intensity with Fe2+ ions was lower than that with Fe3+ ions (LOD 3.5 nM), confirming the complexation
strength and electron transfer process between the d4- orbitals in the excited state. In contrast, the
PDFe3+ (1:2) chelate was very stable and showed high emission intensity, even in the presence of
paramagnetic cations such as Cu2+, Ni2+, Co2+, Ag2+, and Fe2+, as well as Zn2+, Pb2+, and Hg2+ ions (Fig. S17).
The pH-dependent fluorescence studies showed that the designed PD probe worked effectively at
neutral to basic pH (6.5 to 8.5) (Fig. S18). Under strongly acidic conditions, micromolar solutions of PD
were light blue with a high emission intensity, indicating randomly aggregated states and PET inhibition
from the protonated heteroatoms of the diethanolamine units towards the anthracene unit upon
photochemical excitation. Based on coulombic attraction, ionic species can be stabilised/extracted using
another ionic species of opposite charge. Accordingly, such a strategy can be very useful to detect
biologically relevant ions more selectively based on hard–hard and weak–weak (acid/base) interactions.
Thus, in situ generated ensembles were utilised to detect the anions either through displacement or
cooperative binding because of the high probability of coulombic interactions.
19

Fig. 7. Relative fluorescence quenching ratio histogram of the PDFe3+ (1:2) complex with 25.0 eq. of
various tetrabutylammonium anions in ACN:H2O (3:7, v/v%) at ex = 376 nm and em = 429 nm. Error
range ±3% at em = 429 nm.
Accordingly, in the fluorescence studies of the ensembles generated in situ with 25.0 eq. of
-
-
2-
-
3-
tetrabutylammonium (TBA) salts of F-, Cl-, Br-, I-, OAc-, NO3 , HSO4 , SO4 , OBz-, OH-, SCN-, H2PO4 , HP2O7 ,
-
ClO4 , and CN- ions, only CN- ions showed nearly 93% quenching of the anthracene monomer emission
(Fig. 7). The obtained interaction was found to be very selective towards CN- ions, confirming that Fe3+
was selectively extruded from the ensemble in saline buffer. The binding constant was obtained from
the fluorescence titration studies by fitting the data at em = 429 nm through non-linear regression
analysis. The value, Ka = 1.72  108 M-1, indicated that the decomplexation process was stronger than
the complexation process with Fe3+ in buffered media (Fig. 8). The saturation point was obtained after
20

the addition of 12.0 eq. of CN- ions, demonstrating that 6 eq. of CN- ions were required to extrude each
Fe3+ ion from the complex. The obtained LOD for CN- ions was 0.1 nM (Fig. S19).
Fig. 8. Fluorescence titration studies of the PDFe3+ (1:2) complex (0.1 M) with TBACN salt in ACN:H2O
(3:7, v/v%) at pH 7.0 at ex = 376 nm. The inset shows the plot of the emission intensity of complex vs.
eq. of CN- ions at em = 429 nm.
The rapid-decomplexation-aided ‘switch off’ response on the addition of CN- ions was attributed to the
highly stable Fe(CN)3 complex, where additional CN- ions stabilise the complex by association in the
coordination sphere of the Fe3+ ions in saline buffer. In contrast, partial 1H-NMR studies of the PD probe
(1.50 mM) in CD3CN showed dramatic shifts in the proton resonances. Upon replacement with 70% v of
D2O in CD3CN, downfield shifts in the protons labelled a and b were observed. In contrast, aromatic
protons were shifted upfield because of aggregation at millimolar concentrations. Upon addition of 2.0
eq. of Fe(ClO4)3nH2O (anhydrous basis), the aliphatic protons shifted from Ha  3.47 to 4.31 ppm (
0.84 ppm) and Hb d 2.75 to 3.83 ppm ( 1.08 ppm), respectively. In addition, the Hc protons shifted
from  4.73 to 5.57 ppm ( 0.84 ppm). Furthermore, the chemical shifts of the anthracene protons
21

labelled e and d also shifted significantly, from He  7.47 to 7.85 ppm ( 0.38 ppm) and Hd 8.49 to 8.57
ppm ( 0.08 ppm). Upon the addition of 6.0 eq. of TBACN ions, the chemical shifts of the protons
moved upfield (Table 1) with a broadening of PD proton resonances, indicating the loss of the d5-
interactions via the decomplexation process in mixed aqueous media.
Even though we were unable to identify the cation– interactions between the Fe3+ ions and the
1
anthracene unit in the complex, the substantial shifts observed in the H-NMR studies suggested that
the weak d5 electronic coupling through the methylene spacer resulted in weak paramagnetic properties.
1
Fig. 9. Partial H-NMR spectra of (i) PD (1.5 mM) in CD3CN, with (ii) D2O (70% v), (iii) (ii)+Fe3+ (2.0 eq.),
and (iv) (iii) + TBCN.
22

The fluorescence titration studies of PD in the absence of buffer revealed dramatic quenching (Fig. S20)
with a very good association constant, Ka = 1.12  104 M-1. Upon stochiometric validation through Job’s
plot (Fig. S21), FT-IR spectra of PD and the PDKCN (1:6) complex showed shifts with reduced intensities
of the bands corresponding to hydroxyl and diethyl groups, and anthracene units towards lower
wavenumbers (Fig. S22). These results indicate the formation of hydrogen bonds between the hydroxyl
groups and CN- units, which make a contribution to the optical properties of the excited state of the
anthracene units. In addition, partial 1H-NMR studies show the substantial upfield shifts in the aliphatic
protons labelled a and b and downfield shifts in the Hc resonances (Table 1) of PD in CD3CN-D2O (1:3
v/v%), confirming the strength of hydrogen bonding of the probe to CN- anions (Fig. S23).
Table 1. Proton chemical shifts of PD with various eq. of Fe3+ and CN- ions in CD3CN:D2O (2.5:7.5 v/v%).
Protons
Ha
Hb
Hc
Hd
He
PD (CD3CN)
3.46
2.73
2.78
4.71
4.80
8.66
8.41
7.58
7.40
PD (CD3CN-D2O) 3.46
PD + Fe3+
4.28
4.00
5.66
8.57
8.12
PD + Fe3+ + CN-
3.65
0.00
3.43
0.05
4.74
0.09
8.40
7.50
-0.25
-0.18
D2O)
0.82
1.20
0.86
0.16
0.72
(Fe3+)
-0.63
2.55
-0.57
1.25
-0.92
5.09
0.29
-0.17
8.20
-0.62
7.06
(CN-)
PD + CN-
-0.91
-1.53
-0.21
-0.34
(Only CN-)

aChemical shift differences () in ppm. Positive and negative values indicate downfield and upfield
shifts, respectively. Here, (D2O) = PD(CD3CN-D2O)  PD(CD3CN), (Fe3+) = PD + Fe3+ (2.0 eq.)  PD(CD3CN-D2O)
,
(CN-) = PD + Fe3+ (2.0 eq.) + CN- (6.0 eq.)  PD + Fe3+ (2.0 eq.), and (Only CN-) was obtained after the
addition of TBACN alone in CD3CN:D2O (2.5: 7.5 v/v%).
23

Considering the potential utility of the simple PD probe, its sensing abilities were analysed in
tap/drinking water (Table S3), as well as ensemble-coated paper-based strips (Fig. 10) [38,39]. The
surface modification of the paper-based strips with respective analytes (Fe3+/CN-) was characterised by
polarised optical microscopy (Fig. S24 and Fig. 25). The obtained LODs are at the nanomolar level (0.005
and 0.5 ng/cm3 for Fe3+ and CN- ions, respectively), indicating the sensitivity of the solid-state-emission-
aided sensing capabilities on cellulose fibres.
Fig. 10. Probe-coated paper strips showing the fluorescence enhancement upon the addition of various
concentrations of (i) Fe3+ ions and (ii) CN- ions (2.0 L each on Whatman filter paper surface). The
images show the sequential addition of Fe3+ (2 L) and CN- (0.8 L) ions, respectively. Note: All images
were processed using Image J, and the highly accurate intensity-based analyte sensing was further
validated (see reference 30 for details). For clear visualization, high and low amounts of Fe3+ and CN-
ions were added, respectively. Because of the high emission intensity, it was very difficult to obtain high-
resolution images with low concentrations of CN- ions, so only PD was used for coating, as shown in Fig.
(ii).
Colorimetric assays are the most popular methods to detect analytes in the visible light range. PCV is
one of the most widely used colorimetric dyes for the analysis of various ionic species in a buffered
aqueous medium. Upon the addition of 2.0 eq. of PCV to the 0.1 M solution of PD in ACN-H2O (2:8, v/v)
in PBS buffer, a colourless solution was obtained, having UV-Vis electronic absorption peaks at 319, 336,
356, 375, and 400 nm. Following the addition of Fe3+ (0.5, 1.0, 1.5 and 2.0 eq.) ions, broad bands in 400-
525 nm range with increased molar absorptivity at 319, 336, 356, and 400 nm (Fig. S26) were obtained.
24

On the sequential addition of PCV to the solution of the PDFe3+ (1:2) complex, a yellow colour (Fig. 11)
developed, and new bands appeared at 433 and 600 nm along with the hyperchromic shifts at 319, 339,
356, 375, and 400 nm (Fig. 12). The band observed at 600 nm is characteristic of PCVM3+ chelates,
resulting in a blue colour because of the formation of coordination bonds between the catechol units
and M3+ [40]. The blue/yellow colour transitions generally depend on the molar extension coefficient at
600/430 nm range respectively.
Fig. 11. Highly selective PCV-based (4.0 eq.) colorimetric assay of the PD (0.1 M) in ACN:H2O (3:7, v/v%)
at pH 7.0 (PBS buffer) with Fe3+ and CN- ions.
In contrast, the band intensity observed at 433 nm ( = 1.98  106) was higher than that at 600 nm ( =
5.85  105); this is the reason for the yellow colour in buffered ACN medium. The observed yellow colour
became brighter after the addition of 5.0 eq. of PCV. Such unusual behaviour was not observed with
other hard metals such as Al3+, Ga3+, and In3+ (Fig. S27) or other di- and monovalent metals (Fig. 13),
demonstrating that the complexation process is selective towards Fe3+ ions with competitive PD binding
in the presence of PCV. Interestingly, on the addition of TBACN (20 eq.) to the yellow solution of the
complex, a green colour developed, with characteristic bands at 715 (broad), 600, and 415 nm. To
investigate the complexation-aided optical changes, pH-dependent studies with various ions were
carried out (Fig. 14). The bands observed after the sequential addition of PCV, Fe3+, and CN- ions in the
25

presence/absence of PD at different pH ranges (Table S3) suggested that the complexation process is
mainly dependent on the ionic species present in the solution.
Fig. 12. UV-Vis spectra of PD (0.1 M) with Fe3+ (2.0 eq.) and PCV (0 to 5.0 eq.) in ACN-H2O (3:7, v/v%) at
pH 7.0 (PBS buffer). The inset shows the optical changes after the addition of 4.0 eq. of PCV.
Fig. 13. Colorimetric changes upon addition of Fe2+ and Fe3+ ions to PCVPD (0.1 M) solution in the
presence of 5.0 eq. of metal ions (Mn+, where n = 1 and 2) in ACN-H2O (3:7, v/v%) at pH 7.0 in PBS buffer.
In contrast, the colours obtained under different conditions showed no overlap with the colours
obtained with Fe3+/CN- ions at pH 7.0 in the presence of PCV, thus allowing the clear discrimination of
analytes through the colorimetric assay. Further, because of the fluorophoric nature of the PD probe,
26

the emission intensity was also measured under similar stoichiometric conditions as those of UV-Vis
spectroscopy. Interestingly, the fluorogenic changes were observed with high specificity in the regular
order of analytes. The addition of an Fe3+ (2.0 eq.) solution of the PD probe (0.1 M) in ACN:H2O (3:7,
v/v) at pH 7.0 resulted in high fluorescence emission, followed by the addition of 1, 2, and 5 eq. of PCV,
which resulted in ~2%, 13%, and 68% quenching, respectively. The addition of 12.0 eq. of CN- ions to the
just-quenched solution showed >97% quenching and a green colour change. The specific stoichiometric
addition of Fe3+/CN- ions in the presence of PCV showed similar changes in colour in the UV-Vis window,
as well as emissions centred at 427 nm. To evaluate the effects of the pH and solvent conditions further,
UV-Vis studies were performed under various pH and solvent conditions. The observed characteristic
colours (tabulated in Table S3) with sequential addition of Fe3+ (2.0 eq.) and CN- (12.0 eq.) ions were
very specific at pH 7.0 in PBS buffer (Fig. 14), confirming the selective colorimetric response without any
interference. In addition, the effect of dissolved salts on optical response variations in buffered medium
demonstrated that counter anions such as chlorides also significantly affect the complex stability and
optical signals.
27

Fig. 14. UV-Vis spectra of PD (0.1 M) with PCV, Fe3+, and CN- ions under different pH conditions in
ACN:H2O (3:7, v/v%). Note: Spectra were recorded upon the addition of analytes such as Fe3+/CN- along
with PCV ions in exact stoichiometric amounts with respect to the PD probe. To record the spectrum of
pure PCV under different pH conditions in the presence of analytes, 0.5 M of PCV and excesses of 5.0
and 15.0 eq. of Fe3+ and CN- ions, respectively, were used.
Based on the electronic absorption spectral variations with different analytes in the presence of PCV, we
were motivated to study the fluorogenic behaviour. Accordingly, in the presence of PCV, based on the
formation of yellow colour in buffered ACN media, quenching occurred upon excitation at ex = 376 nm.
Such quenching was attributed to resonance energy transfer (RET) from anthracene to PCV. The
emission spectra of the donor (PD, anthracene monomer emission) were overlapped with the
absorption spectra of acceptor PCV), validating this hypothesis (Fig. S28). Fluorescence titration studies
showed saturation at ~4.0 eq., after which the quenching intensity was not steep. The observed KSV was
in the 105 M-1 range (Fig. S29).
With the rapid development of molecular electronics, photonic logic circuits have attracted more
attention in recent years. Accordingly, many chemically driven logic circuits have been designed and
fabricated. However, most of these devices fail to recognise the sub-micromolar-level chemical species
(analytes) under physiological conditions with reversibility. In addition, such logic circuits fail to
28

construct the necessary feedback loops to create the initial conditions for many cycles, resulting in
inefficient data processing in lab-on-a-chip devices. In that sense, design of simple and robust
fluorescent probes for in vitro and in vivo studies with excellent reversibility is crucial for the preparation
of novel biomaterials for a new generation of lab-on-a-chip devices [41-44]. Thus, we designed a simple
and robust INHIBIT logic gate, constructed using Fe3+ and CN- species as an input at pH 7.0 in saline
buffer (Fig. 15). The two chemical inputs (In), Fe3+ and CN-, are labelled as InFe3+ and InCN-, respectively.
The threshold value of fluorescence intensity at the output (426 nm) was set to 400 a.u. When the
fluorescence intensity was higher than this threshold value, the outcome was assigned as ‘1’, and when
it was lower, the response was assigned as ‘0’. These responses define the ‘on and ‘off’ states,
respectively. The reversible and reconfigurable sequences of the set/reset logic operations explain the
memory feature with ‘read-write-erase-read’ functions through the output signal at em = 426 nm. The
set input (InFe3+ = 1) causes the fluorescence at 426 nm to switch on. The system reads the encoded
information as written, and logic operations are saved as output ‘1’. The fluorescence at 426 nm is
switched off by resetting the input (InCN- = 1), which erases the stored encoded information, and logic
operations are saved as output ‘0’. The designed circuit was able to function for more than six cycles (Fig.
S30) with effective feedback looping from the output signals under general physiological conditions.
29