Pyrene Excimer Formation Kinetics

August Rothenberger

Partners: Stephen Nagle and Andrew Cleek

Department of Chemistry, The Pennsylvania State University, University Park, PA 16802

Abstract ...........................................................................................................................................

Using 337.1nm and 336.0nm lasers, the rate constants for the excimer formation kinetics

of pyrene-k0, k1, and k2-were calculated to be 3.10*106±3.1*103 s-1, 2.65*109±2.6*108 s-1M-1, and

2.56*107±8.1*104 s-1, respectively. Deviation from ideal rate constants was determined to be

4.15% by analysis of the k0 value and attributed to the data processing technique.

Introduction ....................................................................................................................................

The polycyclic aromatic molecule pyrene (Py) is commonly used in fluorescence

spectroscopy. Py can absorb radiation and promote an electron to the singlet excited state (*Py).1

At high concentrations of pyrene, *Py molecules can react with a ground state Py to form an

excited state dimer, called the pyrene excimer (*Ex).1 If the initial Py concentration is too low,

no *Ex will form due to limited collisions of *Py and Py molecules.1 The excited state molecules

can relax by fluorescence, each emitting a different wavelength.

𝑃𝑦 + ℎ𝜐 → ∗𝑃𝑦 (1)
∗
𝑃𝑦 → 𝑃𝑦 + ℎ𝜐𝑚 (2)

𝑃𝑦 + ∗𝑃𝑦 ⇌ ∗𝐸𝑥 (3)

∗
𝐸𝑥 → 2𝑃𝑦 + ℎ𝜐𝐸 (4)

In Equation 2, the rate of *Py fluorescence decay is defined by the first order rate

constant k0. The fluorescence decay rate of *Ex in Equation 4 is represented by k2. The second
order equilibrium rate constants for the formation and dissociation of *Ex in Equation 3 are

given by k1 and k-1. If a solution has sufficient Py concentration, k0 and k1 compete with one

another due to *Ex formation and *Py relaxation occurring in solution simultaneously.1

If it is assumed that k-1 is negligible,2 the concentrations of *Py and *Ex are defined by:1

𝑑[ ∗𝑃𝑦]⁄ ∗
𝑑𝑡 = −(𝑘0 + 𝑘1 [𝑃𝑦])[ 𝑃𝑦] (5)

𝑑[ ∗𝐸𝑥]⁄ ∗ ∗
𝑑𝑡 = −(𝑘2 )[ 𝐸𝑥] + 𝑘1 [𝑃𝑦][ 𝑃𝑦] (6)

The condition of pseudo-first order kinetics can be assumed in samples with high enough Py

concentration since there will be more *Ex than *Py in solution. If the initial concentration of

*Py is written as [*Py]0 and the initial concentration of *Ex is assumed to be zero, Equations 5

and 6 can be solved and rewritten as:1

[ ∗𝑃𝑦] = [ ∗𝑃𝑦]0 ∗ 𝑒 −𝑘𝑜𝑏𝑠 ∗𝑡 (7)

𝑘1 [𝑃𝑦][ ∗𝑃𝑦]0
[ ∗𝐸𝑥 ] = ( ⁄𝑘 − 𝑘 ) (𝑒 −𝑘𝑜𝑏𝑠 ∗𝑡 − 𝑒 −𝑘2 ∗𝑡 ) (8)
2 𝑜𝑏𝑠

kobs is the observed rate constant for the luminescence decay, accounting for both *Py

fluorescence and *Ex formation.1 It is defined as:

𝑘𝑜𝑏𝑠 = 𝑘0 + 𝑘1 [𝑃𝑦] (9)

Experimental ..................................................................................................................................

Solutions of 10µM, 0.5mM, 1.0mM. and 1.5mM pyrene in decane were prepared in glass

cuvettes and purged with nitrogen gas. The absorbance of each sample was recorded from 420 to

300nm using a Cary 400 FL-3 UV-Vis spectrophotometer. The fluorescence intensity for the

samples was measured at 335nm for the 10µM sample and 355nm for all other samples using a

Fluorolog FL-3 spectrofluorometer.

 The fluorescence decay was measured at 337.1nm for the 10µM sample using an SRS

NL-100 laser with pulse width of 3.5ns. For all other samples, the decay was measured at 366nm

using an SRS OBB laser with pulse width of 0.8ns. Fluorescent radiation was collected by an

Electro-Optics Technology model 23-2618A photodiode at a rise time of 0.5ns and filtered

through a 400±5nm optical interference filter and 500nm band-pass optical filter. Intensity was

recorded using a Tektronix model TDS 2022B 200MHz storage oscilloscope for the SRS NL-

100 and a Tektronix model TDS 3022B 500MHz storage oscilloscope for SRS OBB.

Results and Discussion...................................................................................................................

Determination of the Emission Wavelength

The absorbance spectra for each concentration were plotted in Figure 1. The most

prominent absorption peak was determined for each sample. For the 10µM sample, the peak lies

at 337nm. However, the absorbance is too high at this wavelength for the other samples, instead

the most prominent peak is at 366nm.

 337 nm 366 nm

Figure 1. Absorption Spectra for Pyrene Samples

Due to this discrepancy in absorbance, the two sets of samples cannot be excited at the

same wavelength. When the 10µM sample is excited with a laser at 337.1nm, the beam excites

all pyrene molecules through its path in solution, as shown in Figure 2. However, if a 0.5mM

sample is placed in front of the same laser, the light is absorbed by the solution before it is able

to pass through the entire sample, as seen in Figure 3. The high absorbance prevents the

formation of a homogeneous distribution of excited state pyrene molecules. In this case the

excimer also cannot form homogeneously in the cuvette, so it is impossible to properly measure

the kinetics of the system. Exciting the 0.5, 1.0, and 1.5mM samples with a wavelength of

336.0nm rectifies the absorbance issue and allows for excitation throughout the entire solution.
 Figure 2. 10µM at 337.1nm Figure 3. 0.5mM at 337.1nm

The fluorescence of all samples were measured at the specified wavelengths and

normalized to the third vibrational emission peak. The overlaid emission spectra are shown in

Figure 4. Pyrene fluorescence is shown from 360-410nm, while the excimer fluorescence is

shown in the region of 430-500nm. The excimer fluorescence intensity increases as a function of

pyrene concentration since more excimer molecules are able to form at higher concentrations of

excited pyrene.

Figure 4. Fluorescence emission spectra normalized to the third vibrational peak

Calculation of kobs

The 10µM sample was excited at 337.1nm and the luminescence decay measured through

a 400±5nm optical interference filter. The raw data was processed by removing data points

before the maximum initial intensity and after the relaxation had completed. The emission

intensity was normalized to its maximum value and plotted as Figure 5. The normalization was

performed by dividing all intensities by the maximum intensity.

𝐼𝑛𝑡𝑒𝑛𝑠𝑖𝑡𝑦
𝐼𝑛𝑡𝑒𝑛𝑠𝑖𝑡𝑦𝑁𝑜𝑟𝑚𝑎𝑙𝑖𝑧𝑒𝑑 = ⁄𝐼𝑛𝑡𝑒𝑛𝑠𝑖𝑡𝑦 (10)
𝑀𝑎𝑥

An exponential regression was used to fit the data.

Figure 5. Normalized luminescence intensity decay for the 10µM pyrene sample

The intensity can be linearized by taking the natural logarithm of the normalized intensity values.

The linearization is performed to Equation 7, which becomes:

ln[ ∗𝑃𝑦] = ln[ ∗𝑃𝑦]0 − 𝑘𝑜𝑏𝑠 ∗ 𝑡 (11)

The linearized data is plotted against time in Figure 6.

 Figure 6. Natural log of normalized intensity versus time for 10µM sample

From Equation 10, the slope of the line of best fit is equal to -kobs. Since there is almost

no *Ex formation due to the low concentration of Py, this value of kobs is the same as k0 for

10µM sample.1 A linear regression was performed using the LINEST equation in Excel at a 95%

confidence interval to return more precise values.

Table 1: Linear Regression of Figure 6

kobs (s-1) -3.10E+06 2.73E-02 b
Δkobs (s-1) 3.09E+03 1.61E-03 Δb
R^2 9.98E-01 3.78E-02 s(y)

The k0 value for all pyrene excitations can thus be determined to be:

𝑘0 = 3.10 ∗ 106 ± 3.09 ∗ 103 𝑠 −1

The luminescence intensity for the 0.5, 1.0, and 1.5mM pyrene samples were measured at

366.0nm through a 400±5nm optical interference filter. The emission intensity was normalized

and plotted as against time in Figure 7. The natural logarithm of the luminescence intensities was

plotted against time in Figure 8.

 Figure 7. Normalized luminescence intensity decay for the (mM) pyrene samples

Figure 8. Natural log of normalized intensity versus time for (mM) samples

A linear regression was performed in the same manner as Table 1 to calculate the kobs values for

the three samples. The rate constants and uncertainties are provided in Table 2.
 Table 2: Linear Regression of Figure 8
0.5mM 1.0mM 1.5mM
kobs 3.89E+06 5.67E+06 6.89E+06
Δkobs 4.34E+03 7.89E+03 1.33E+04
R^2 9.96E-01 9.94E-01 9.88E-01

Calculation of k1

Since kobs is defined by equation (9), k1 can be determined by plotting the values of kobs

versus pyrene concentration. The values from Table 1 and Table 2 were used to construct Figure

9 below. The slope of the data is equal to k1 and the intercept is the model’s calculation of k0.

k(obs) versus [Pyrene]

0.0E+00 5.0E-04 1.0E-03 1.5E-03
8.00E+06
7.00E+06
6.00E+06
k(obs) (s^-1)

5.00E+06
4.00E+06
y = 3E+09x + 3E+06
3.00E+06 R² = 0.9818
2.00E+06
1.00E+06
0.00E+00
[Pyrene] (M)

Figure 9. kobs versus [Py] to calculate k1

A linear regression was performed and the results provided in Table 3.

Table 3: Linear Regression of Figure 9

k1 (s M-1)
-1
2.65E+09 2.89E+06 k0 (s-1)
Δk1 (s-1M-1) 2.55E+08 2.39E+05 Δk0 (s-1)
R^2 0.981754 2.84E+05 s(y)

The discrepancy between the k0 calculated above and the value calculated in Table 1 is

due to the imprecise truncation of emission data required to obtain a logarithmic fit.
 k-1 represents the rate constant of the reverse reaction in Equation 3: the conversion of

*Ex to Py and *Py. This process is most likely to occur due to collision of the excimer with

solvent molecules.2 However, the binding energy of the excimer is sufficiently high and its

lifetime sufficiently short that it is very unlikely that the molecule will decay before fluorescing.2

Therefore, k-1 can be neglected for kinetic calculations.

Calculation of k2

The laser pulse profile for the 0.5, 1.0, and 1.5mM pyrene samples excited with a

336.0nm laser and filtered through a 500nm band-pass optical filter were plotted against the

luminescence build up in units of time. The results are shown in Figure 10.

Figure 10. Laser pulse profile for the (mM) samples

To determine k2, the plots in Figure 10 must be fit to Equation 8. Solving for the rate

constant would require multivariable calculus, so it is best solved using a mathematical fitting

software. A MATLAB script was provided to fit the data, the resulting plots are displayed in

Figures 11, 12, and 13 below.

Figure 11. Laser pulse profile and luminescence build up for 0.5mM sample

Figure 12. Laser pulse profile and luminescence build up for 1.0mM sample
 Figure 13. Laser pulse profile and luminescence build up for 1.5mM sample

The results are summarized in Table 4.

Table 4: Summary of k2 Values

0.5mM 1.0mM 1.5mM
k2 2.27E+07 2.59E+07 2.83E+07
Δk2 6.12E+04 3.85E+04 3.72E+04
R^2 0.9985 0.9996 0.9997

k2 should be constant for pyrene excimer formation, so the results were averaged to obtain a

single value. The rate constants for pyrene excitation are summarized below.

𝑘0 = 3.10 ∗ 106 ± 3.1 ∗ 103 𝑠 −1

𝑘1 = 2.65 ∗ 109 ± 2.6 ∗ 108 𝑀−1 𝑠 −1

𝑘2 = 2.56 ∗ 107 ± 8.1 ∗ 104 𝑠 −1

Deviation from Standard Behavior

Deviation in the model used to calculate the rate constants can be determined using

Figure 9. It was assumed that the 10µM pyrene sample’s kobs was equal to k0 due to negligible

*Ex formation. Since the y-intercept of the plot should equal the 10µM point, the deviation can

be calculated by:

|2.89E + 06 − 3.10E + 06|⁄

%𝐷𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 = 2.89E + 06 ∗ 100% = 4.15%

This deviation in our model is due to the data processing required for the emission spectra to

obtain a fit by exponential regression. The truncation of data after relaxation was determined

qualitatively, causing slightly different kobs values for the samples than would be ideal for a

perfect linear fit in Figure 9.

Conclusion ......................................................................................................................................

Rate constants for the pyrene excimer formation kinetics were calculated to be

3.10*106±3.1*103 s-1, 2.65*109±2.6*108 s-1M-1, and 2.56*107±8.1*104 s-1. These values show

4.15% deviation from the ideal, which is attributed to qualitative data processing techniques

performed to apply a logarithmic fit for the emission spectra.

Sample Calculations.......................................................................................................................

A normalization using Equation 10 is performed below for one of the data points used to

create Figure 5.

𝐼𝑛𝑡𝑒𝑛𝑠𝑖𝑡𝑦𝑁 = 0.123887⁄0.129223 = 0.958702

The error propagation for the averaged k2 valued calculated from Table 4 is shown below.

∆𝑘2 = √(6.12 ∗ 104 )2 + (3.85 ∗ 104 )2 +(3.72 ∗ 104 )2 = 8.13 ∗ 104 𝑠 −1

Acknowledgements ........................................................................................................................

Reagents, materials, and laboratory space were provided by the Penn State Chemistry

Department. Special thanks to Rebecca Katz and Prof. Bratoljub H. Milosavljevic for assistance

and guidance in writing this report.

References .......................................................................................................................................

1. Milosavljevic, B. “Lab Packet for CHEM 457 Experimental Physical Chemistry.” Fall 2018.

2. Hanlon, A.D. and Milosavljevic, B.H., Photochem. Photobiol. Sci., 12 (5), 787-797 (2013).