Luminescence Quenching and Upconversion Experiment

CH454 Physical Chemistry Lab #3

Introduction

This laboratory experiment will provide you with an introduction to the study the quantum yield
and kinetics of luminescence quenching. One of the most studied molecules in this field is the
Ru(bpy)32+ complex. This complex is used as a photosensitizer in photoelectrochemistry (solar
energy) applications. It is also a nice laboratory example, because it has a long-lived excited
state. In order to understand the observations in this experiment you will want to review first
order and second order kinetics, fluorescence and phosphorescence, and various quenching
mechanisms. It is important to differentiate between energy and the charge transfer mechanisms
of quenching. Here will we will use energy transfer quenching combined with a second process
of triplet-triplet annihilation in the acceptor to make a higher energy singlet state.

The donor for triplet energy transfer used in this experiment is Ru(bpy)32+, which is shown in
Figure 1.

Figure 1. The structure of the Ru(bpy)32+ ion

These bidentate ligands form tris complexes with Ru(II) that have approximately
D3 point group symmetry. The electronic absorption spectrum of Ru(bpy)32+ at room temperature
consists of a broad band from 400-460 nm. This transition consists of a (spin allowed) metal-to-
ligand charge transfer (MLCT) from a molecular orbital that has mostly Ru(II) d orbital character
to a molecular orbital that has mostly ligand π orbital character. Since the band is a MLCT band
it has a large excited state dipole moment.

Ru(II) complexes are excited by metal-to-ligand charge transfer (MLCT) transitions. For
example, in the bipy complex the process initially forms a singlet MLCT state. However, this
singlet state forms a triplet state by intersystem crossing.
1
Ru(bpy)32+ → 3
Ru(bpy)32+
1
 One reason for this is the heavy atom effect of
Ru, which increases the spin-orbit coupling process
leading to a spin flip. Once the triplet state is formed the
excited state of 3Ru(bpy)32+ can be quenched by
anthracene. Since Ru(bpy)32+ has evolved to the triplet
state the luminescence is phosphorescence rather than
fluorescence. However, many people mistakenly refer to
the luminescence as fluorescence. In this experiment we
will use 9,10-diphenylanthracene (DPA) since that
molecule cannot undergo photochemical crosslinking
reactions.
3
Ru(bpy)32+ + DPA → Ru(bpy)32+ + 3DPA

This aspect of the experiment is fairly routine

since phosphorescence quenching of Ru(bpy)32+ obeys
standard quenching kinetics, so-called Stern-Volmer
kinetics. The Stern-Volmer model is the same for
fluorescence and phosphorescence quenching. The
experiment we will study has an added process, which is the combination of two 3DPA
molecules by triplet-triplet annihilation to make a higher energy singlet DPA.

2 3DPA → DPA + 1DPA

When these reactions are combined they are given the name “upconversion” to refer to the fact
that the output photon from the fluorescence of 1DPA has a higher energy than the input photon
used to excite Ru(bipy)32+. The issues surrounding the quenching of phosphorescence are the
core of the lab. The observation of upconversion is mainly to show you one of the numerous
applications of quenching processes.

In this experiment you will measure both the relative quantum yield and the
photochemistry of electronically excited state of the tris(2,2'-bipyridine) ruthenium (II) ion
[Ru(bpy)3]2+ in which Ru(II) ion is surrounded by three bidentate chelating ligands of bipyridine
(bpy) as shown in figure below:

Fluorescence vs. Phosphorescence. Excited states can be deactivated either by be radiative

or nonradiative processes. Spin allowed radiative transitions are usually short-lived (< 10−8 s)
and give rise to luminescence, which is known as fluorescence. Spin forbidden radiative
transitions can also be observed as a long-lived luminescence, which is called phosphorescence.
Since singlet-triplet transitions are spin-forbidden, triplet excited states can be form only by
means of a process known as intersystem crossing. The intersystem designation refers to the
fact that the molecular system crosses from the singlet to the triplet state manifold. Although
phosphorescence is often difficult to observe in many organic molecules, it is readily observed in
Ru(bipy)32+.

2
Kinetics of luminescence and quenching. Ru(II) diimine complexes are relatively highly
luminescent, even though observed emission is from a triplet 3MLCT excited state. The reason
for that is in a high spin-orbit coupling (Ru is a heavy metal) which couples singlet and triplet
manifold of states in so called 'configurational interaction' and results in relatively short radiative
lifetime. At room temperature and in aqueous solution, 3Ru(bpy)32+ shows strong luminescence
at about 600 nm with lifetime around τo ~ 600 ns. There are two processes that compete with the
phosphorescent deactivation. These are internal non-radiative process intrinsic to 3Ru(bpy)32+ and
quenching processes. In this case we will focus on energy transfer quenching, which means that
the triplet state of 3Ru(bpy)32+ is transferred to the energy transfer acceptor, which is 1,9-
diphenylanthracene (DPA).

Luminescence: 3
Ru(bpy)32+ → Ru(bpy)32+ + hv (1)

Nonradiative deactivation: 3
Ru(bpy)32+ → Ru(bpy)32+ + heat (2)

Quenching: 3
Ru(bpy)32+ + Q → Ru(bpy)32+ + 3Q (3)

We will use a nitrogen laser which provides excitation pulses of about 4 nsec half-width. A pulse
excites a dilute solution of [Ru(bpy)3]2+ and produces a population of excited states
*[Ru(bpy)3]2+. Without a quencher, this population decays radiatively with an observed rate
constant of kobs = kr + knr, where the radiative rate constant is kr and the non-radiative rate
constant is knr. If we assume that the laser pulse produces a population No of 3[Ru(bpy)3]2+ at
time t = 0, the population can be shown to decay exponentially (first order kinetics), resulting in
the following time dependence of the emission intensity I (photons/sec):

𝐼𝐼 = 𝑘𝑘𝑟𝑟 [ 3𝑅𝑅𝑅𝑅(𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏)2+
3 ] = 𝑘𝑘𝑟𝑟 𝑁𝑁0 exp{−𝑡𝑡/𝜏𝜏0 } (4)

where

𝜏𝜏0 = (𝑘𝑘𝑟𝑟 + 𝑘𝑘𝑛𝑛𝑛𝑛 )−1 (5)

The quenched emission decay will proceed with an accelerated speed. In other words, its lifetime
will be shorter:

𝐼𝐼 = 𝑘𝑘𝑟𝑟 𝑁𝑁0 exp{−𝑡𝑡/𝜏𝜏} (6)

Where

𝜏𝜏 = (𝑘𝑘𝑟𝑟 + 𝑘𝑘𝑛𝑛𝑛𝑛 + 𝑘𝑘𝑞𝑞 [𝑄𝑄])−1 (7)

3
Thus, the rate constant for the quenching reaction, kq, can be obtained by analyzing the
luminescence data. The quenching rate constant, kq, is determined from the so-called Stern-
Volmer relation:

𝜏𝜏0
= 1 + 𝜏𝜏0 𝑘𝑘𝑞𝑞 [𝑄𝑄] (8)
𝜏𝜏

which can be obtained from the rearrangement of Equations 5 and 7. The quantum yield of
luminescence is proportional to the luminescence intensity, so that the Stern-Volmer relation
may also be written as:

𝐼𝐼0
= 1 + 𝜏𝜏0 𝑘𝑘𝑞𝑞 [𝑄𝑄] (9)
𝐼𝐼

where Io and τo are the luminescence intensity and the luminescence lifetime of 3Ru(bpy)32+ in
the absence of quencher.

The equations above suggests that kq may be obtained from an experiment where the
luminescence lifetime or the luminescence intensity of Ru(bpy)32+ is measured at different
quencher concentrations. A plot of τo/τ (or Ιo/Ι) versus [Q] is expected to be linear, with a slope:

𝐾𝐾𝑠𝑠𝑠𝑠 = 𝜏𝜏0 𝑘𝑘𝑞𝑞 (10)

where Ksv is the Stern-Volmer rate constant.

Stern-Volmer quenching is derived with the assumption of dynamic quenching, which means
that the quenching is free to diffuse in solution and that it interacts with the luminescent
molecule by a second-order process. However, one must always be aware of the possibility that
two molecules have formed a complex and therefore that the quenching is static, rather than
dynamic. In order to be sure that a process is due to dynamic quenching it is necessary to
determine both the yield and the lifetime.

Diffusion limited reactions. In a dynamic process the quenching rate constant kq is a

bimolecular rate constant (has units M-1s-1) and describes a process, in which the reactants first
have to encounter and then react. Such reactions are often referred to as diffusion-
assisted reactions. It is convenient to separate a reaction into two steps. First, there is a diffusion
controlled formation of intermediate complex (A*...Q) or transition state, with kdiff, and second
the transition state can either react with the (first order) rate constant kp or dissociate without
reacting with the rate constant k-diff (also first order) :

kdiff kp
A* + Q <=> (Α∗...Q) −−> Products (11)
k-diff

Then the overall rate constant k' can be written as:

4
 𝑘𝑘𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 𝑘𝑘𝑝𝑝
𝑘𝑘 ′ = (12)
𝑘𝑘𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 + 𝑘𝑘𝑝𝑝

which in the limit of fast reaction rate kp >> k-diff simplifies to:

𝑘𝑘 ′ = 𝑘𝑘𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 (13)

when the diffusion becomes a limiting step. This case would correspond to the diffusion
controlled (or diffusion limited) regime, when no matter how fast kp is, the observed k' would be
defined by how fast molecules encounter due to diffusion. The value of kdiff obviously depends
on the diffusion coefficients of both reactants:

𝐷𝐷 = 𝐷𝐷𝐴𝐴 + 𝐷𝐷𝑄𝑄 (14)

According to the rate equation:

𝑘𝑘𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 = 4𝜋𝜋𝑁𝑁𝑎𝑎 𝐷𝐷𝑅𝑅𝑞𝑞 (15)

where Na is the Avogadro's number 6.022 x 1023 mol-1, and Rq is the separation at which the
quenching takes place, also called the quenching radius. For neutral molecules, the quenching
radius equals the sum of molecular radii:

𝑅𝑅𝑞𝑞 = 𝑅𝑅𝐴𝐴 + 𝑅𝑅𝑄𝑄 (16)

but the situation can differ dramatically if both species are charged. The Coulombic interaction
leads to the following condition:

Rq > RA + RQ - for oppositely charged ions (17)

Rq < RA + RQ - for ions of the same charge (18)

Mechanisms of quenching. A variety of inorganic and organic species may act as quenchers.
Determining kq does not give any insight into the mechanism of quenching. For a given
system, kq may reflect a combination of energy and electron transfer processes. Energy transfer
could occur in either the singlet or the triplet manifold. By detailed study of the system and
application of control experiments one can determine whether the mechanism involves energy
transfer or electron transfer. Triplet energy transfer can be established in part using the
sensitivity of the process to O2. Since O2 has a triplet ground state it is an effective triplet
quencher (see below). When applied to simple systems, these criteria help determine the
predominant mechanism of quenching.

Energy transfer: *Ru(bpy)32+ + Q → Ru(bpy)32+ + Q* (19)

Reductive electron transfer: *Ru(bpy)32+ + Q → Ru(bpy)3+ + Q+ (20)

5
 Oxidative electron transfer: *Ru(bpy)32+ + Q → Ru(bpy)33+ + Q− (21)

This laboratory experiment will provide an example of energy transfer quenching. When energy
transfer takes place according to a transition dipole-transition dipole mechanism it is known as
Förster energy transfer. Efficient energy transfer depends on a number of factors. There is a 1/R6
distance dependence for Förster energy transfer. This distance dependence arises from the
dipolar nature of the process since the electric field that results from a dipole decays at 1/R3. A
pair of dipoles that interact mutually has a distance dependence of (1/R3)2 = 1/R6. There is also
an orientation dependence for Förster energy transfer since the probability of energy transferring
from one dipole to another depends on the square of the cosine of the angle between the two
transition dipole moments. Clearly, there are many orientations in solution and one must carry
out an orientation average. For a given distance, energy transfer depends on the Franck-Condon
factors of the two molecules. The luminescence of the donor must have overlap with the
absorption of acceptor. This condition does not mean that light is emitted by the donor and then
absorbed by the acceptor. Indeed, energy transfer is NOT such a process, but it is a coherent
oscillation of the two transition dipole moments subject to the correction conditions of
orientation and energetics. One of your tasks in this laboratory will be to compare the emission
of the donor (3Ru(bipy)32+) with the absorption spectrum of the acceptor (DPA).

Degassing of samples. Given the role played by O2 as a triplet quencher it is important to degas
the solution of any sample, in which one wishes to measure the triplet energy transfer yield or
rate constant. Degassing can be accomplished by two methods. First, one can use freeze-pump-
thaw as a method. Second, one can bubble a gas such as N2 or Ar through the sample for a period
of time (usually about 20 minutes). Freeze-pump-thaw is considered the best way to degas a
sample. However, it is quite time consuming. One usually freezes the sample in liquid N2 or
some other cryogen, then pumps on the sample by attaching it to a vacuum manifold and finally
permits the sample to thaw after it is sealed so it has no contact with the atmosphere. Usually this
process is repeated three times to ensure that all of the residual O2 gas is removed from the
sample. Degassing by bubbling a solvent is much faster, but is not as well controlled and has the
disadvantage that one will often lose a fair amount of volatile solvent due to evaporation. One
can control for evaporation by weighing the sample and either recording the new concentration
or adding a small amount of (degassed!) solvent to the sample using a gas tight syringe. In
general, all transfers in such experiments can be carried out using so-called Hamilton syringes.
Hamilton has been the leading company that makes gas tight syringes for many years. Standard
Hamilton syringes range in volume from 10 µL to 1000 µL.

CAUTION: When degassing using N2 or Ar gas bubbled through the sample always ensure that
there are two syringe needles, an input and output to prevent a pressure build-up in the sample.

Experimental

This laboratory experiment consists of two parts, 1. The measurement of the emission spectrum
and 2. The measurement of the emission lifetime of Ru(bipy)32+ in solution with added DPA
quencher molecules. You need to make solutions of both Ru(bipy)32+and DPA in
dichloromethane (DCM). The ideal solution for measurement of the phosphorescence quenching

6
will have an absorbance of ~ 0.2 in the MLCT band of Ru(bipy)32+ at ~ 452 nm. However, for
observation of upconversion a higher concentration is desirable such that the A ~ 1.0 at 452 nm.
Thus, the experiment should be run at two different concentrations of Ru(bipy)32+. To determine
the appropriate concentrations you may use the extinction coefficient of 14,600 M−1 cm−1 at λmax
= 452 nm. Note that we will use the PF6− salt with the chemical formula (PF6)2Ru(bipy)3 with a
formula weight of 859.6 a.m.u. The solutions of DPA will need to be much more concentrated
since you will need to inject small volumes of this solution into the solution of (PF6)2Ru(bipy)3
in order to study a series of different concentrations of the quencher DPA (formula weight 330.4
a.m.u.). Using this background you may prepare the sample using the following step-by-step
procedure:

1. The volume required in the quartz cuvette for phosphorescence and time-resolved
measurements is ~3 mL.
2. This solution must be degassed. This is done by bubbling N2 gas through the cuvette for
20 minutes. The cuvette has a septum cap so that it should be air tight. When bubbling
N2 always make sure to place a second needle in the septum as an outlet so that pressure
does not build up inside.
3. You will need to add aliquots of DPA of volumes ranging from 10-100 µL to this
solution. The stock solution of DPA should be approximately 20 mM. Add aliquots of
this solution successively to the cuvette and acquire an emission spectrum and a time-
resolved kinetics trace for each concentration. Use at least 5 concentrations of DPA (in
addition to the pure Ru(bipy)32+ sample) ranging from 0 – 500 nM.
4. It is important to obtain the absorption spectrum of each solution since the absorption
spectrum can be used to correct for any change in concentration in Ru(bipy)32+ due to the
addition of DPA.
5. Each sample will be studied by two experiments:
A. Fluorescence emission from 550 to 850 nm. You will use the Edinburgh Instruments
fluorescence spectrometer. The instructions are given below.
B. Time-resolved kinetics using the nanosecond fluorescence spectrometer. The
instructions are given below.
6. Save the data in each experiment in ACSII format. For example, on the Edinburgh
Instruments spectrometer use the EXPORT selection on the menu to write the data to
disk. The data on the time-resolved nanosecond kinetics apparatus are stored on a
Tektronix fast oscilloscope. The data need to be saved to a thumb drive of < 2Gb.

CAUTION #1: Xe arc lamps output a high voltage spike when the initially powered on. It is a
good idea to power on the Xe arc lamp before other more sensitive circuits. Computers are
fortunately not sensitive to these spikes.

CAUTION #2: Photomultiplier tubes (PMT) are sensitive detectors that can be destroyed by
high light levels. Please follow the guidelines of each instrument closely and never change
settings without consulting with your instructor. DO NOT CHANGE THE PMT VOLTAGE
SETINGS!

Edinburgh Instruments Fluorometer

7
 There are three power supplies that need to be switched on. These are on the rack above
the laser table. These consist of the Xe-arc lamp (always turn this on first!), then the
photomultiplier tube (PMT) power supply and the temperature controller (in any order after the
lamp). There is also a power switch on other side for the internal controls of the fluorometer.
This switch is sometimes left on.

The software on the computer is FL900. Open the software and check the signal at the
initial wavelength settings. There are two digital meters, one of the reference and one for the
sample PMT. Monitor the number of counts per second recorded by the PMT. The bandwidth
(Dl) will mostly be set to 4.00 nm for both the excitation and the emission. However, it is a good
idea to set it to a smaller value (for example, 1.00 nm) when you first make a measurement. The
idea is to limit the number of photons striking the PMT. This number should be less than 8 x 106.

Set the excitation wavelength to 470 nm. Set the emission scan from 550 – 850 nm. The
dwell time should be 0.1 seconds and the increment 1 nm. Set the number of scans to 3. The data
will be written to a binary data file on the hard drive. In order to be able to use the data for
analysis you will need to export it. In scientific software the export function is the usual way to
write data in ASCII format as columns of numbers separate by either a comma or tab. This
format can be read by programs such as Excel, Origin, Igor, etc. that are used for data fitting and
analysis.

Time-Resolved Emission Apparatus

There is a black notebook on the table describing the steps to setup a time-resolved scan.
You will work with your instructor to turn on the instrument and set up to take data. You may
take data on your own by switching out samples and recording data, but do not change settings
for the PMT or laser. You may examine the settings on the digital oscilloscope and your
instructor will show you how to check for the trigger level, the settings in terms of time scale,
voltage scale and so on. These are described in the book as well.

The Upconversion Measurement

Upconversion is an idea that has been developed and demonstrated in the Castellano
research group. We have chosen a quenching experiment that added functionality. We can
attempt to estimate the yield of upconversion by measuring the fluorescence emission of 1DPA
created by triplet-triplet annihilation. You measure that emission between 250 and 400 nm.
However, may need a concentrated sample to see the effect.

Report

You will acquire three kinds of data in this laboratory experiment, 1. UV-vis absorption
spectra (on the HP8453 photodiode array), 2. Phosphorescence spectra from 550 – 850 nm on the
EI fluorometer and 3. Time-resolved emission kinetics. You will use the Stern-Volmer relation
to make a linear plot of the relative integrated phosphorescence yield as a function of [Q], the
quenching concentration. From a straight-line fit of these data (using linear least squares) you
will obtain a slope. However, to further analyze that slope you need an independent measure of

8
the observed phosphorescence decay time from the time-resolved setup. If the quenching is
dynamic quenching then you should also see a change in the lifetime as you increase the
quencher concentration. Thus, the time-resolved experiment is a check on the assumptions of the
Stern-Volmer derivation (dynamic quenching).

In the analysis you will first need to normalize the data to the initial Ru(bipy)32+
concentration using the UV-vis spectra. As you add aliquots of DPA the absorbance at 452 nm
should decrease slightly. Thus, you can multiply the fluorescence by the value Ainitial/Asample for
each sample you study (Ainitial means the absorbance of pure Ru(bipy)32+ at 452 nm and Asample
refers to the same measurement on each sample studied. Keep careful notes and a simple
consistent nomenclature to avoid confusion in the interpretation of the files.

Integrate each phosphorescence signal in a software package such as Excel, Origin or

Igor. Fit each decay to an exponential model in a software package. Examine the residuals to
ensure that the fit to a single exponential is adequate. Use the integrated values obtained from the
experiment to determine the Stern-Volmer rate constant (Eqn. 9). You will then need to use the
measured lifetime, 𝜏𝜏0 , to determine the quenching rate constant (Eqn. 10).

Try to estimate the quantum yield for production of the upconversion signal in one
sample. Clearly you will need to add enough DPA to completely quench the 3Ru(bipy)32+
emission.

Points for discussion

Look in the chemical literature and find diffusion coefficients for the molecules studied
here or similar molecules. Using an estimate of the diffusion coefficient calculate the quenching
radius. Discuss the assumptions that go into this calculation (as discussed above).

Compare the Ru(bipy)32+ emission spectrum to the DPA absorption spectrum. Keep in
mind the criteria for energy transfer and discuss the features and relationship of these two
spectra. Discuss the appropriateness of this pair of molecules for triplet energy transfer.

What are the consequences of incomplete deoxygenation of the sample. Which

observables are the most sensitive to O2 quenching. Include the production of 1O2 in your
discussion.

References

1. Wilke B.M. and Castellano, F.N. “Photochemical Upconversion: A Physical or

Inorganic Chemistry Experiment for Undergraduates Using a Conventional
Fluorimeter” J. Chem. Ed. 2013, 90, 786-789
2. Singh-Rachford, T.N.;Castellano, F. N. "Photon upconversion based on sensitized
triplet-triplet annihilation" Coord Chem. Rev. 2010, 254, 2560-2573
3. Stern, O; Volmer, M "The Fading Time of Fluorescence" Phys. Zetischr. 1919, 20,
183-188

9
 4. Wilkinson F.; Helman, W.R.; Ross A.B. "Quantum Yields for the Photosensitized
Formation of the Lowest Electronically Excited Singlet State of Molecular Oxygen in
Solution" J. Phys. Chem. Ref. Data 1993, 22, 113-150

Appendix. A discussion of electron transfer quenching of the Ru(bipy)32+ excited state.

The electron transfer rate constant, kET, depends:

(i) distance between the donor and acceptor,
(ii) the degree of quantum mechanical coupling between the molecular orbitals of donor and
acceptor, and
(iii) the free energy change, ∆G°
(iv) the reorganization energy, λ.
The latter parameter refers to the energy cost incurred by molecular rearrangements that must
result from the transfer of charge along a finite distance. Generally, it is expected that rates of
electron transfer will increase with decreasing donor-acceptor distances, and that the maximum
rate will be observed when the reaction is activationless, i.e., when ∆G° is negative, preferably
when ∆G° = - λ. In other words, the rate is optimized when the standard free energy change for
the reaction is matched exactly by the energy required for reorganization of the donor, acceptor,
and solvent molecules. In the table below you see the redox potentials for the species involved in
this experiment. By analysing it you should conclude what quenching mechanisms are plausible
for different quenchers. NB. The exact expression for

Species Redox potential

Ru(bpy)33+ + e- --> Ru(bpy)32+ + 1.26 V
Ru(bpy)32+ + e- --> Ru(bpy)3+ +1.28 V
*Ru(bpy)32+ --> Ru(bpy)33+ + e- + 0.84 V
Ru(bpy)32+ --> *Ru(bpy)32+ (+ 2.1 V)
3- - 4-
Fe(CN)6 + e --> Fe(CN)6 + 0.36 V
3+ - 2+
Fe(H2O)6 + e --> Fe(H2O)6 + 0.77 V
O2 + e---> O2- − 0.365 V

In principle, there is sufficient potential energy in a photo-generated redox pair that one
devise a scheme to split water into H2 and O2. The goal of light-driven water splitting has been
approached by hundreds of different research teams. The processes relevant to the process using
Ru(bipy)32+ are summarized below, where L represents the diimine ligand and Q represents the
oxidant:

RuL32+ + hν → *RuL32+
*RuL32+ + Q → RuL33+ + Q-

10
 2 RuL33+ + H2O → 2 RuL32+ + 1/2 O2 + 2H+
Q- + H+ → Q + 1/2 H2

In other words, solar energy can be is used to make fuels from water. Although the chemistry of
such a solar cell is straightforward, there are some technological barriers to be overcome before
ruthenium (II) complexes can form the basis of a commercial photovoltaic device.

Other applications for photooxidation of ruthenium (II) complexes involves development of due-
sensitized photovoltaic devices, where *RuL32+ is oxydized by a network of TiO2 nanoparticles
and consequently RuL33+ is reduced with a help of I3−/I− redox pair. Resulting charges recombine
at the electrodes, thus, producing a phovoltage equal (without a load) to the overall redox
potential through all the steps.

11