Red-to-Blue Triplet–Triplet Annihilation Upconversion for Calcium Sensing

Triplet–triplet annihilation upconversion is a bimolecular process converting low-energy photons into high-energy photons. Here, we report a calcium-sensing system working via triplet–triplet annihilation (TTA) upconverted emission. The probe itself was obtained by covalent conjugation of a blue emitter, perylene, with a calcium-chelating moiety, and it was sensitized by the red-light-absorbing photosensitizer palladium(II) tetraphenyltetrabenzoporphyrin (PdTPTBP). Sensing was selective for Ca2+ and occurred in the micromolar domain. In deoxygenated conditions, the TTA upconverted luminescence gradually appeared upon adding an increasing concentration of calcium ions, to reach a maximum upconversion quantum yield of 0.0020.


General Methods
Reagents were purchased from Sigma Aldrich, Fisher Scientific, FluoroChem, VWR and TCI.Solvents were used directly if not mentioned otherwise.Dry DMF was obtained as a result of distillation under nitrogen atmosphere followed by settling over activated 3 Å molecular sieves under nitrogen atmosphese for two days.Thin-layer chromatography (TLC) was performed with Merk silica-coated aluminum plates (F60 with F254 indicator).Column chromatography was performed using 60 Å silica gel (0.04 -0.063 mm, Screening Devices B.V.).NMR spectra were recorded on a Brucker AV-400, Brucker AV-I-500 or Brucker AV-III-600 spectrometer at 298 K. Recorded spectra were analysed with MestReNova software.Electrospray ionization mass spectra (ESI-MS) were recorded with a Thermo Fisher MSQ Plus mass spectrometer with 17-2000 m/z detection range.High resolution mass spectrometry (HRMS) was performed with a Thermo Finnigan LTQ Orbitap with electrospray ionization (ESI) method.Elemental Analysis was performed by the Microanalytical Laboratory Kolbe (Oberhausen, Germany).Analytical LC-MS was made with an analytical NUCLEODUR™ C18 Gravity column (3 µm, 50 x 4.6 mm,; Macherey-Nagel) connected to a Vanquish™ UHPLC system with a Vanquish™ Diode Array detector coupled to a LCQ™ Fleet via ESI (all Thermo Fisher Scientific).
1 H-NMR titration was performed with Brucker AV-400 at room temperature.The recorded spectra were analysed with MestReNova software.
1 H-DOSY experiments were made with AV-III-600 at 298 K in 3 mm EPR tubes with delays d20 = 90 ms and p30 = 480 μs.Obtained data were processed with Brucker TopSpin software.Diffusion coefficients determined with the Stejskal-Tanner equation fitting of T1/T2 relaxation module analysis of integrated signals.Binding constants were determined with isothermal titration calorimetry (ITC) using a MicroCal VP-ITC.Titration was performed at 20 °C with 28 injections (2 μL first and 10 μL following).Milli-Q water was used as a reference, reference power was set for 10 μcal/s, initial delay 60 s, stirring speed 307 rpm.Obtained data were processed with MicroCal ITC-ORIGIN Analysis Software.
Fluorescence emission spectra were recorded with HORIBA Aqualog in 1 cm quarz cuvette.Recorded spectra were processed with Origin 2022 software.All another custom-built setups for upconversion measurements described in corresponding sections below.

3-perylene boronic ester (5)
In a 500 mL two-necked round-bottom flask equipped with stirring bar and a reflux condenser, 3-perylenebromide (1.81 g, 5.46 mmol) as a solid mixture with perylene, potassium acetate (1.56 g, 16.4 mmol), and bis(pinacolato)diboron (3.86 g, 15.3 mmol) were dissolved in 1,4dioxane (200 mL) and purged with nitrogen.Then 1,1'-bis(diphenylphosphino)ferrocene-palladium(II) dichloride (280 mg, 0.383 mmol) was added to the reaction mixture.The reaction mixture was stirred at 70 °C under nitrogen atmosphere overnight.After cooling to room temperature, the solvent was removed under reduced pressure.The residue was extracted with dichloromethane (1x200 mL), washed with sat.NaHCO 3 (aq.)(2x200 mL) and brine (1x200 mL).The organic layer was dried over magnesium sulfate, filtered, and dried under reduced pressure.The product was isolated by column chromatography on silica gel using a dichloromethane / hexane (v/v = 1:1) eluent giving   (6)   In 50 mL round-bottom flask equipped with stirring bar and a Din-Stark cap, 3-perylene boronic ester 5 (373 mg, 0.99 mmol) and Br-BAPTA-Et 4 3 (330 mg, 0.49 mmol) were dissolved in toluene (16 mL) and ethanol (3 mL).The solution was purged with nitrogen during 30 min.Then, Pd(PPh 3 ) 4 (57 mg, 0.05 mmol) was added to the solution followed by 2 M aq.K 2 CO 3 (0.5 mL).The reaction was stirred at 85 o C degrees for 20 min.Afterwards the temperature was raised to 100 o C and the mixture was stirred for 22 h more.After cooling to room temperature, the reaction mixture was washed with water (3x30 mL) and the organic layer was dried over anhydrous Na 2 SO 4 .After evaporation, ethanol (5 mL) was added to the residue and the precipitate was filtered, washed with ethanol (2x5 mL) and ether (2x5 mL).The crude product was purified by silica gel column chromatography using DCM/EtOAc/triethylamine (v/v, 9/1/0.01)as eluent, followed by preparative TLC in the same eluent to give 66 mg of the title compound 6 (0.493 mmol, 15 %).

In methanol
In methanol two solutions of the sensor 1 (100 μM) as a titrated sample and CaCl 2 •2H 2 O (1 mM) as titrating solution were prepared.The samples were degassed in MicroCal ThermoVac during 3 minutes under vacuum at 19 o C prior to measurements.Titration was performed at 20 o C with 28 injections (2 μL first followed by 10 μL).Methanol was used as a reference, reference power was set for 10 μcal/s, initial delay 60 s, stirring speed 307 rpm.The obtained data were processed with MicroCal ITC-ORIGIN Analysis Software.In figure S.4.2 we observe several processes upon increasing the calcium concentration.We assume that from 0.0 to 0.5 molar guest:host ratio binding is clearly an entropically-driven process assigned to binding to calcium of 1 to 1 ratio.After 1:1 host:guest ratio two enthalpy-driven processes seem to occur.

Fluorescence quantum yield measurements
Fluorescence quantum yield (QY) was estimated by comparison with perylene fluorescence QY in cyclohexene (0.94), 4 and calculated by equation S.5.1.: Where  f is a fluorescence QY of the molecule of interest;  fR is a fluorescence QY of the reference; F is the integrated fluorescence intensity, OD is the optical density, and n is the refractive index of the solvent.Solutions of the sensor 1 in methanol and perylene in cyclohexane were prepared in such a concentration that its OD at 400 nm wavelength was below 0.05.Emission spectra were recorded with HORIBA Aqualog spectrometer in a 1 cm quarz cuvette at 21 o C. Recorded spectra were processed with Origin 2022 software.Competition study: Solutions of metal cations with 6 μM of 1 were prepared in methanol, after which Ca 2+ (1 μM) was added, samples were irradiated with 400 nm light.Control is the ΔF/F 0 of the sensor in the presence of 1 μM of Ca 2+ .

Phosphorescence quenching in presence of the annihilator 1
Steady-state emission spectra were recorded with the same setup described above.635 nm laser (16.2 mW, 2.9 mm 2 , 559 mW/cm 2 ) was used as an irradiation source.All measurements were made at 24 o C temperature.In in 0.25 x 1 cm (optical path 1 cm) quarz cuvette equipped with a septum a solution of PdTPTBP (2.5 μM) in methanol was prepared and deoxygenated by means of purging with argon during 20 min.After addition of a portion of the annihilator 1 solution the sample was further deoxygenated during 10 min.The recorded spectra were processed with AvaSoft® and OriginPro 2022® software (  Pd-TPTBP phosphorescence quenching in presence of different concentrations of annihilator 1. Conditions: λ ex = 635 nm, laser power = 16.2 mW, laser beam cross-section area = 2.9 mm 2 ; power density = 559 mW/cm 2 , methanol; 24 o C. Insert shows a zoom of the 400-600 nm emission region.

Titration of the PS and the sensor 1 solution in methanol in deoxygenated conditions with Ca 2+ at r.t.
Emission spectra were recorded with the setup described above.635 nm laser (16.2 mW, 2.9 mm 2 , 559 mW/cm 2 ) was used as an irradiation source.All measurements were made at 24 o C temperature.In in 0.25 x 1 cm (optical path 0.675 cm) quarz cuvette equipped with a septum a solution of PdTPTBP (2.5 μM) and the sensor 1 (75 μM) in methanol was prepared and deoxygenated by means of purging with argon during 20 min.Separately, a solution of CaCl 2 •2H 2 O (1 mM) was prepared in methanol.After addition of a portion of the CaCl 2 •2H 2 O solution, the sample was further deoxygenated during 10 min.The recorded spectra were proceeded with AvaSoft® and OriginPro 2022® software.

Control titration of the PS and perylene solution in methanol in deoxygenated conditions with Ca 2+ at r.t.
Emission spectra were recorded with the setup described above.635 nm laser (16.2 mW, 2.9 mm 2 , 559 mW/cm 2 ) was used as an irradiation source.All measurements were made at 24 o C temperature.In in 0.25 x 1 cm (optical path 0.675 cm) quarz cuvette equipped with a septum a solution of PdTPTBP ( 2  Titration of PdTPTBP (2.5 μM) and perylene (75 μM) solution in methanol in deoxygenated conditions with Ca 2+ .Conditions: λ ex = 635 nm, laser power = 16.2 mW, laser beam cross-section area = 2.9 mm 2 ; power density = 559 mW/cm 2 , methanol; 24 o C.

Quantum yield measurements by absolute method
The measurement was performed with a custom-build setup (Figure S.5.5).The procedure for quantum yield determination had been adapted from the work previously made in our group. 6The luminescence spectrum of a blank sample with pure methanol was measured first.For that sample, an OD 2.0 neutral density filter was placed in the filter holder 6 to dim the laser light and protect the sensor.Then a sample containing Pd-TPTBP (2.5 μM), sensor 1 (25 μM or 75 μM) and CaCl where λ 1 -λ 2 -wavelengths range of upconverted signal; λ 3 -λ 4 -wavelengths range of laser signal luminescence for the blank and the TTA-UC samples, T SP625 (λ) -transmittance of the 625 nm short pass filter, T OD2.0 (λ) -transmittance of the neutral density filter with OD = 2.0.Laser signal of the blank and TTA-UC samples and its upconverted emission.Conditions: λ ex = 635 nm, laser power = 16.2 mW, laser beam cross-section area = 2.9 mm 2 ; power density = 559 mW/cm 2 ; solvent: methanol; temperature: 24 o C., argon.
The recorded upconverted emission spectra and laser signal for the blank sample and the TTA-UC sample were corrected with transmission of 625 nm short-pass and OD 2.0 neutral density filters, respectively.Obtained spectra were integrated in the ranges λ 1 -λ 2 = 440 nm -625 nm and λ 3 -λ 4 = 625 nm -649 nm.The final upconversion quantum yield was calculated according to Equation S.  I th of PS (2.5 μM) and 1 (25 μM) system determination in the presence of 75 μM of Ca 2+ in methanol.b) Quantum yield dependence of power density for PS (2.5 μM) and 1 (25 μM and 75 μM).Conditions: λ ex = 635 nm laser beam cross-section area = 5.7 mm 2 ; solvent: methanol; temperature: 24 o C, argon.

Beam profiling
A combination of CinCam CMOS-1201-Nano with neutral density filter OD 2.0 from Edmund, placed between the camera and the laser beam source was used for cross section area measurement.Beam diameters and cross section area were determined with RayCiLite software.The 10% laser beam diameter were determined by Gauss Distribution fitting in RayCiLite software (Figure S.5.6).The obtained cross section area in mm 2 was used for power density calculation in parts 5.4 and 5.5.

Fluorescence lifetime measurements
A combination of a Tsunami® Ultrafast Ti:Sapphire Osscilator, Hamamatsu® HPDTA streak camera C4334, a Newport Spectra-Physics® pulse selector (model 3980) and a CHROMEX spectrograph was used for recording of fluorescence decay spectra 7 .Ones were recorded in deaerated and aerated conditions in methanol with the 420 -565 nm spectral range and 10 ns time window.Pulses centered at 400 nm were used for sample excitation.Instrumental response factor was recorded with a glass reflecting the excitation beam into the detector.Obtained spectra processed with Origin 2022® software.Samples with the sensor 1 (6 µM) and different Ca 2+ concentrations (0 µM; 0.5 µM; 2.5 µM; 1 µM; 7 µM and 10 µM) were freshly prepared in deoxygenated conditions in 1x1 cm fluorescence quartz cuvettes equipped with Schlenk line connectors with valves.For recording spectra in aerated conditions, the valves were left open for 20 min and the samples were used the second time.

Nanosecond transient absorption spectroscopy
Transient absorption spectra were recorded with a custom-built setup described previously 8 .Samples were excited with 627 nm laser pump (The Continuum® Surelite TM Nd:YAG laser in combination with The Continuum® OPO Plus) with the power of 200 µJ.A 75 W Xenon arc lamp was used as a probe.610 nm short-pass or 640 nm long-pass filters were placed before the detector to avoid the pump-scattering.Measurements were made in 1 cm fluorescence quartz cuvettes modified with a Schlenk connecting adapter.Stability of the samples were checked before and after the measurement with recording of their Uv-vis spectra.Recorded spectra were processed with Origin 2022® software.For Stern-Volmer plot, different samples containing Pd-TPTBP (5 µM) and different sensor 1 concentrations (0 µM; 25 µM; 75 µM and 125 µM) in deaerated methanol were freshly prepared.The spectra were recorded in the range 350 -800 nm.For the Stern-Volmer plot lifetimes t 1 were taken (Figure S.6.2, 4b).Pd-TPTBP (5 μM) -Sensor 1 (75 μM) excited state absorption decay at 500 nm wavelength in presence of different Ca 2+ concentrations.We don't observe change in biexponential decay lifetimes in presence of different calcium cation concentrations.

Computational studies
Geometry optimized models of sensor 1 with 0, 1, or 2 Ca 2+ cations were constructed and optimized using ADF modeling suite from SCM. 9 Conformer searches were performed for each of the molecules using OpenBabel's genetic algorithm. 10The resulting 20 conformers per sensor with 0, 1, and 2 Ca 2+ ions bound were geometry optimized using TeraChem 11 using density functional theory (DFT: rev PBE0/6-311++g) 12 and COSMO to simulate solvation in water.In order to evaluate the total energy of the two binding event and evaluate whether the formation of the 1:2 complex [1•2Ca•6H 2 O] is entropically favorable, we considered the following 2-step balanced reaction.4 sodium ions were added to keep the total charge of the system to zero.12 water molecules were included in the model to stabilize the Ca 2+ and Na + ions and simulate their solvation when not bound or partially bound to sensor 1.The multiplication coefficients are reported in table S.

Figure S. 4 . 1 .
Figure S.4.1.Isothermal calorimetry data of sensor 1 titration with CaCl 2 •2H 2 O in 10 mM HEPES buffer at 20 o C. a) ITC data for titration; b) reference data for buffer titration with the same Ca 2+ solution.
.5 μM) and perylene (75 μM) in methanol was prepared and deoxygenated by means of purging with argon during 20 min.Separately, a solution of CaCl 2 •2H 2 O (1 mM) was prepared in methanol.After addition of a portion of the CaCl 2 •2H 2 O solution, the sample was further deoxygenated during 10 min.The recorded spectra were proceeded with AvaSoft® and OriginPro 2022® software (Figure S.5.4).
2 x2H 2 O (75 μM) in methanol was degassed by purging argon during 30 min.A luminescence spectrum of this sample was then recorded.The neutral S18 density filter was replaced by a 625 nm short-pass filter, and the upconverted luminescence spectra was recorded.The obtained data were processed with OriginPro 2022® software (Figure S.5.6) using the equation defining the upconversion quantum yield (Equation S.5.2):
-em is the emitted photon flux of the upconverted luminescence [photons/s] and q p-abs is the absorbed photon flux by the photosensitizer [photons/s].

Figure S. 5 . 1 .
Figure S.5.1.Example of beam profiling data obtained with RayCiLite.a) 2D cross-section of laser beam; b) X Cross section of the beam; c) Y cross section of the beam; d) result of fitting performed in the software.

Figure S. 6 . 1 .
Figure S.6.1.Normalized fluorescence decay of sensor 1 in presence of different Ca 2+ concentrations (a) and the plot of the ratio of singlet state lifetimes of sensor 1 and it's fluorescence intensities ratio versus Ca 2+ concentration (b).Here we observe a static interaction of sensor 1 with Ca 2+ .

S5 1 .4. Per-BAPTA-Et 4
was added to the reaction mixture, which was stirred vigorously at room temperature.Reaction was monitored by an-HPLC.As soon as the peak of the starting compound had disappeared, the reaction was stopped, first by rotary evaporation of the organic solvents, then freeze-drying the water phase.200 mg (81 %) of the target compound was obtained as a tetrasodium salt containing 1 eq. of NaOH.This compound was pure enough according to HPLC (15 min linear gradient from 10% CH 3 CN/0.1%TFAaq. to 90% CH 3 CN/0.1%TFAaq.), to be used directly.It must be kept cold and it is sensitive to acids, which decompose it by decarboxylation.

Table S .5.1. Fluorescence quantum yields for 1 in presence of different Ca 2+ concentrations calculated with equation S.5.1.
[a] Reference sample in cyclohexane with QY 0.94.