Flash photolysis is a technique to generate and study excited states. The excited states are generated by a high energy pulse of light and monitored by changes in absorbance of the monitoring beam passing through the sample.
The technique of flash photolysis was introduced by Norrish and Porter in 1949.(Norrish and Porter, 1949; Porter, 1950; Porter and Wilkinson, 1961) In these early works the sample was excited with light from a discharge lamp, and monitored either photographically over many wavelengths by the use of a rapid weaker pulse following at a given time post excitation, or photo electrically at a single wavelength by a continuous lamp as a function of time. The relatively long duration of the exciting flash restricted the study of transient species to those with very long lifetimes because monitoring cannot begin until the light pulse has finished. This is known as the 'blind time' of the system. To study shorter lived species, excitation sources with much shorter pulse duration and faster detection systems were required.
With the introduction of nanosecond pulsed lasers to this technique, laser flash photolysis provides an extremely powerful tool for the study of very short lived species. Indeed, with the arrival of picosecond and femtosecond lasers, and the improvement in monitoring systems, it is now possible to use laser flash photolysis to study processes occurring in the singlet state. In this case the laser provides both the excitation, and monitoring source. However, in most nanosecond experimental arrangements the monitoring source is a xenon arc lamp. This limits the time resolution to between 0.1µs and 1µs.
The principle of flash photolysis, and laser flash photolysis depends on the assumption that the excited state in question absorbs a different amount of light to the ground state. If the excited state absorbs more light than the ground state, less light reaches the detection system, and a deflection occurs equal to the reduction in the light. This is known as a transient absorption and is associated with areas of low ground state absorption. Conversely, if the excited state absorbs less light than the ground state, more light reaches the detection system, and a deflection occurs equal to the increase in the light. This is known as a transient depletion, and is associated with areas of high ground state absorption.
The experimental arrangement for nanosecond laser flash photolysis, constructed and developed during my PhD is shown in figure 2.1.
Essentially laser flash photolysis apparatus may be regarded as a time resolved absorption spectrophotometer that may be operated in two distinct modes, continuous, or pulsed lamp, depending on the type of work being carried out, and the lifetime of the transient species concerned. The normal mode of operation, and the simplest is the use of a continuous lamp, in which the monitoring source is focused through the sample onto the monochromator slits, and the current from the photomultiplier tube converted to a voltage with 500 ½, 1000 ½, or 10, 000 ½ load resistors. The advantage of using a continuous lamp and large load resistors is that it is possible to use lower light intensities and therefore reduce the risk of photobleaching the sample, or photoconverting it to another species with the monitoring lamp, before the laser fires. In this case one would not be monitoring the excited state of sample, but the photoproduct of the sample. The main disadvantage with the use of a continuous lamp is that with large load resistors the time resolution of the apparatus is reduced, and so a compromise between time resolution, light intensity and load resistor must be sought to give an acceptable signal to noise ratio.
Where a transient is very short lived (t < 5µs) the effect of large load resistors would be to artificially increase the lifetime and therefore give false kinetics. In these cases the monitoring lamp is pulsed to increase its output by over 100 times, and a load resistor of 50½ applied. The length of this pulse is normally only 2.5ms, as this gives a good compromise between baseline stability and increased intensity. Unlike with the continuous lamp where the trigger point of the laser and scope are not critical, in pulsed mode all the events must be timed to occur, in sequence, within the 2.5ms duration of the lamp pulse. The sequence of timing events for operation of laser flash photolysis using pulsed lamp following depression of the fire button is as follows, and is summarised in figure 2.2:
Lamp shutter opens
Arc lamp pulser, pulses monitoring lamp
Laser flash lamps triggered
Back-off triggers and holds the output from the photomultiplier tube at the flat portion on the lamp profile. It then generates an equal and opposite voltage to make the trace appear at zero on the oscilloscope. The back-off displays this voltage as the value Vo.
The back-off then triggers the oscilloscope
The laser pockel cell opens and releases the laser pulse 840µs (ruby laser), or 610µs (Nd:YAG laser) after the initial flashlamp trigger.
Lamp shutter closes.
In the experimental setup used in our laboratory, the transient species were monitored with a 250W Kratos air-cooled Xenon arc lamp (principally operated in pulsed mode) focused through the sample cell. The transmitted light being focused on to the slits of an applied photophysics monochromator (ƒ=3.4) and detected with a Hamamatsu R928 photomultiplier (200-900 nm range). The current output from the photomultiplier was fed through an automatic back-off unit to channel 1 of a Tektronix 2432A digital oscilloscope (sampling rate: 5ns per point, 1024 points ) and terminated with a 50½ load resistor. The displayed transient was then fed to a personal computer for data analysis.
A more recent development in the experimental setup in our laboratory has been the introduction of new software, optical triggering, and the use of continuous lamp instead of pulsed lamp. In this arrangement the triggering is no longer controlled by the timing generator, instead the laser is set to internally trigger at a repetition rate of about 1Hz. The optical trigger is connected to the EXT1 trigger of the oscilloscope, a fibre optic with its head mounted in the laser such that it receives a 10% reflection of the 1064nm fundamental from a glass slide mounted in the beam path, delivers light to the photodiode within the trigger unit. When correctly aligned the the trigger unit can generate a 10V signal with a rise time of a few nanoseconds. The data analysis software continuously checks for a trigger at the oscilloscope. When a trigger is detected the contents of the scope CH1 memory are transferred through a GPIB interface to the computer's memory, this continues until the pre-set number of acquisitions has been made. This software automatically collects data for lamp-on with and without laser, and lamp off with and without laser, the benefit of this being that the software corrects for both baseline non-linearity, and for fluorescence emission, displaying the transient on screen in ÆA and not mV as with the older system.
Laser flash photolysis (and pulse radiolysis) monitors the change in light transmission through a sample, as a function of time, by the change in voltage reaching the oscilloscope. The parameter determined is the change in the sample's absorbance, ÆA, and is determined as follows:
If I1 is assigned to the intensity of the light transmitted through the sample before pulsed excitation (proportional to Vo), and I2 to the intensity at time, t, after pulsed excitation (proportional to Vo-x). By the Beer-Lambert law we can assign an absorbance before, and after excitation; hence by subtraction, the difference absorption ÆA.
Where ÆeT is the triplet difference molar absorption coefficient, [3M*] is the triplet state concentration, and l is the path length. From this we see that the change in absorbance at time, t, is given by ÆA=log{Vo/(Vo-x)}. By determining ÆA for a range of wavelengths, it is possible to plot the change in the transient (usually triplet) difference spectrum as a function of time.
i) First order decays:
In the absence of quencher, a sample excited to the triplet state, 3M*, with a low intensity pulse (< 10% conversion to the triplet state), will decay by radiationless collisional deactivation due to interaction with solvent molecules. This gives a unimolecular (first order) decay rate constant, k1, according to:
Where [3M*]o is the initial concentration of 3M*.
Since the difference in absorbance of triplet and ground state, ÆAT, is directly proportional to [3M*], it follows:
Where 
is the initial difference
absorbance. A plot of lnÆAT versus time,
t, gives a straight line with slope, -k1, and
intercept ln
.
ii) second order decays (quencher induced):
Where the triplet state is being quenched by a second species such as molecular oxygen, O2(3S), an enhanced rate of decay of the triplet state is observed relative to unimolecular decay. This gives an overall rate of decay of [3M*] as:
Normally the ground state oxygen concentration (typically 10-3-10-4M) is much greater than the triplet state (typically 10-6-10-7M) and so pseudo-first order kinetics are observed.
The bimolecular quenching rate constant, kq can be determined from a plot of k2 versus quencher concentration [O2(3S)]. The magnitude of the value of kq is limited by the bimolecular rate constant for diffusion, kd, and is solvent dependent (usually kq ² 3x1010 M-1s-1), and also by spin statistical factors (see section 1.5.3. Singlet molecular oxygen, generation).
iii) Second order decays (triplet-triplet annihilation):
At high laser energies, a significant concentration of triplets can be generated. This increases the probability of triplet collisions causing triplet-triplet annihilation.
Products may be dimers, two ground state species or a combination of S1 and So, the formation of S1 states leading to P-type delayed fluorescence. Under conditions of high laser energies the decay of [3M*] may be given by:
Hence, a plot of lnÆAT versus time, t, enables the determination of dlnÆAT/dt to be made from tangents drawn to the curve. The values of lnÆAT may then be plotted against ÆAT to give kq/ÆeTl from the slope, and k1 from the intercept. This treatment is relatively complex and is rarely used in the routine analysis of transients, It is however important because at high percentage conversions of the ground state to the triplet state, triplet-triplet annihilation will occur giving mixed order kinetics, and so laser pulse energy is usually adjusted to give less than 10% conversion.
If one assumes that following pulsed laser excitation the only species absorbing light at wavelength, l, is the ground state (1Mo) and the lowest excited triplet state (3M*) then we can make a statement about the absorption of light, before, and after the pulse according to the Beer-Lambert law, as follows.
If the triplet difference spectrum has been determined at a constant time after the laser pulse the difference molar absorption coefficient Æeat a wavelength, l, may be determined if the concentration of triplets is known. The triplet difference spectrum can be converted to the triplet-triplet spectrum by addition of the ground state absorption coefficient e, to triplet difference extinction coefficient Æeat each wavelength.
There are three ways to determine the value for Æe, all of which seek to determine the concentration of the triplets formed.
i) Complete conversion
The principle of this method is to increase the laser energy to the point where the ground state is completely converted to the triplet state. At this point only triplet states exist, therefore the concentration of the triplet state [3M*] will be equal to the initial ground state concentration, [1Mo]. ÆAT will be at its maximum obtainable value.
Therefore, ÆeT can be determined from:
The method of complete conversion is shown in figure 2.3 and involves plotting ÆAT as a function of laser energy, taking the asymptote as ÆA. If the ground state concentration is known, Æemay be determined.
The assumption that complete conversion will occur may not always be valid for the following reasons.
Non-uniform sample excitation due to high ground state concentrations resulting in non-linear absorption of excitation light.
Molecule has a low triplet quantum yield (FT).
A short triplet lifetime (tT). If the triplet state is decaying almost as fast as it is being formed (kisc k1), rapid repopulation of the ground state will occur and complete conversion will be kinetically unobtainable.
A long singlet lifetime (tS) compared to the duration of the laser pulse, G, then complete conversion may not be achieved. It was suggested by Carmichael and Hug that for ³95% conversion to occur the following statement must be obeyed.
For a free base porphyrin the singlet state lifetime is Å10ns, and the triplet quantum yield is Å0.8. With a 16ns pulse duration complete conversion may not be achieved. However for anthracene with a singlet lifetime of Å4.9ns and a triplet quantum yield of Å0.71, complete conversion should be achieved.
For the reasons discussed complete conversion may not be achieved, or gives a less than potential value for ÆA, and hence gives a lower limit value for Æe.
ii) Singlet depletion.
This method depends upon measuring the difference absorption in a region where the ground state absorbs. There are several variations on this method depending on the relationship between eT and eS. The simplest method is when the difference absorption is measured in a region where eT « eS (such as porphyrin Soret bands or the Qy band of chlorins and phthalocyanines) (Hadlee and Keller, 1969). In this case at a particular wavelength, ÆAT would be negative, corresponding to a depletion in the triplet difference spectrum, giving.
The triplet concentration can be determined if it is assumed that eT << eS, such that eT may be neglected, giving.
From this [3M*] can be calculated and ÆeT solved for other wavelengths. However, because the assumption that the triplet absorption is negligible at a particular wavelength may not be valid this method yields a lower estimate for Æe.
The second and more complex method involves the assumption that eT varies linearly with wavelength over the region associated with the ground state absorption bands, and is achieved by monitoring different absorption wavelengths around the maximum of the triplet difference spectrum (at the tail of the Soret band).
If ÆAis measured at a constant wavelength interval then equation 2.22 gives equation 2.23.
By the subtractions (2.19-2.20) and (2.20-2.21), two simultaneous equations are obtained, which may be solved if eis known at each wavelength.
In practice, such assumptions as to the behaviour of eT may not always be valid. As a result ÆeT and eT values determined using this method should be treated as an upper limit.
iii) Energy transfer
This method involves the sensitisation of an acceptor triplet state (such as all-trans b-carotene) by a donor triplet state (such as a porphyrin). In an ideal experiment, only the donor triplet is formed by pulsed laser excitation and the individual molecules would have long triplet lifetimes. Hence, it is assumed that by suitably adjusting the solute concentrations, every donor triplet state is quenched to give an acceptor triplet state (see equations 2.26 and 2.27).
Providing that wavelengths are available where only the acceptor triplet absorbs, lAcceptor, and where only the donor triplet absorbs, l Donor, and that the value of ÆeT for one of the two is known at the corresponding wavelength, then the value of ÆeT, of the other can be calculated by re-arrangement of equation 2.27. In practice however, these conditions are seldom ideal and several corrections must be applied.
Absorption of the laser light by the acceptor. If the triplet yield of the acceptor is very small, then the absorption can be corrected for, by measuring the donor absorption on its own, before addition of the acceptor.
Decay of the donor triplet not leading to energy transfer. Kinetic corrections for the decay of triplet donor by means other than energy transfer must be applied.
when k1 k2, the observed ÆAT of the acceptor must be corrected,
Rapid decay of the acceptor. Where the decay of 3A* approaches the rate of its formation, i.e. k3 Å k3,
These correction factors are reduced by measuring ÆAT of donor and acceptor where there is little T1 Æ Tn spectral overlap and employing high concentrations of A. This method gives an upper limit for ÆeT.
In this study FT has been determined by the comparative method, a second method developed by Medinger and Wilkinson (Medinger and Wilkinson, 1965) is applied to strongly fluorescent molecules and depends upon enhanced intersystem crossing (EISC) by heavy atoms such as iodine and bromine, or paramagnetic species such as ground state oxygen. This method involves measuring the decrease in fluorescence intensity, and the associated increase in ÆAT.
The comparative method for determination of triplet state quantum yields involves the use of a reference of precisely known FT and ÆeT at a given wavelength. The method also requires that ÆeT or eT values for the unknown sample are precisely known before FT can be determined:
The triplet state quantum yield (FT) for the sample is given by,
Where ÆA is the difference absorbances of the sample and the reference, (1-10-A) is the sample absorbance at the excitation wavelength, and Io is the laser energy.
The reference compound used in these studied was anthracene in cyclohexane which has ÆeT = 64700 M-1cm-1at 422nm and FT = 0.71(Bensasson et al, 1978) for 355 nm excitation. This method assumes that the ground state absorbance characteristics of the sample do not change over the duration of the excitation pulse. As a result, the laser is attenuated such that depletion of <10% occurs.
Unlike nanosecond laser flash photolysis, which uses a Xenon arc lamp as the monitoring source, picosecond laser flash photolysis (picosecond time-resolved absorption) uses picosecond pulsed lasers as both excitation and monitoring sources.
This method enables the study of events with lifetimes less than 100ps.
In the experimental arrangement (figure 2.4), the 373nm pump beam (excitation) is provided by the frequency doubled 746nm line from a titanium sapphire laser, and the 436nm probe beam (monitoring) is obtained from the frequency doubled mixture of the 1064 line from a Nd:YAG laser and the 746nm line from the titanium sapphire. The high time resolution, and low pulse energy of this equipment prevents monitoring of the complete decay of the species from one pulse as is normal with nanosecond laser flash photolysis. Instead, the decay is built by signal averaging the absorbance of the probe beam at different delays following the pump beam. The delay between the pump beam and the probe beam is introduced with an optical delay line, this works by extending the distance the laser pulse travels and therefore the time it takes to reach the sample. So that the probe beam monitors the area of the sample excited by the pump beam they must overlap exactly. However, as the laser beams have diameters of less than one tenth of a millimetre their alignment is monitored using a CCD camera (see figure 2.4).
Whilst the time resolution of the equipment used in the Laser Support Facility, at the Rutherford Appleton laboratories is very high, the spectral resolution is not and so it is principally used to study kinetics of excited state species. The mathematical interpretation of kinetics used is the same as that used for the study of nanosecond laser flash photolysis.
In pulse radiolysis, the excitation source is a pulse of high energy electrons as described in section 1.3. All experiments were undertaken at the Paterson Laboratories, Manchester, the experimental layout for which is given in figure 2.5. The radiation source at the Paterson Laboratories is a Vickers L-band microwave linear accelerator (LINAC). The accelerator, sample chamber, and monitoring lamp is enclosed in a concrete walled room to prevent leakage. The light from the monitoring lamp is focused through the sample chamber and collimated to a beam that is then taken out of the accelerator room by a series of planar front surface mirrors to the detection system. The operator and the electronics are enclosed in a Faraday Cage on the outside of the accelerator room to reduce electrical noise arising from the electron pulse. The detection system setup is very similar to that used for laser flash photolysis, using a Bausch and Lomb monochromator with interchangeable gratings and photomultipliers/photodiodes. The transient decay is stored on a Tektronix 7612D digitizer interfaced to a Hewlett-Packard 986S computer. Transient analysis for pulse radiolysis is very much the same as for laser flash photolysis. As with laser flash photolysis the procedure is dependent on a change in the light transmission through the sample as a function of time. The difference between the light transmission through the sample before and after the electron pulse is given by ÆA=log{Vo/(Vo-x)} equation 2.1-2.4.
By measuring the difference absorbance at different times, as a function of wavelength it is possible to create a difference spectrum for the excited state and by calculation of extinction coefficients, the difference spectrum may be calculated, in a similar manner to laser flash photolysis.
Singlet oxygen, as described in section 1.5 may be generated by a electron exchange energy transfer (type II reaction) from the triplet state of a sensitiser (figure 2.6) (Foote, 1987; Foote, 1988). The singlet oxygen may then react with a substrate to give oxidation products, or be physically deactivated; or it may decay radiatively, emitting light in the near infrared region at ~1580nm, and ~1270nm. The triply forbidden emission (O2(1Æg)(v~=0) Æ 3S) at 1270nm provides a convenient, accurate and direct method of determining the lifetime, quantum yield, and rate of reaction of singlet oxygen with substrates. The emission at 1580nm is due to (O2(1Æg)(v~=0) Æ 3S&endash;) and is much weaker.
Photosensitised singlet oxygen luminescence at 1270nm was first studied spectroscopically by Krasnovsky (Krasnovsky, 1979) using a modified photomultiplier arrangement. Photomultipliers have poor sensitivity in this region of the spectrum, and have now been replaced with fast, infrared sensitive germanium photodiodes. Recently, ultra-sensitive liquid nitrogen cooled germanium detectors, having sensitivities ~1000 times greater than room temperature systems have been introduced. This, coupled with adaptation of single photon counting techniques (Egorov et al, 1989) and the introduction of the use of high resolution fourier transform interferometry (Bell, 1972; Biggs et al, 1990; Biggs et al, 1987) has led to significant improvements in the sensitivity, and versatility of this method of studying singlet oxygen. The improvements in the sensitivity have enabled detection of singlet oxygen luminescence in biological environments, although the low radiative rates, short singlet oxygen lifetimes, and low oxygen concentrations in biological environments, still make detection of singlet oxygen luminescence in such environments extremely difficult.(Baker and Kanofsky, 1991; Kanofsky, 1989; Kanofsky et al, 1988; Parker, 1987; Rodgers, 1988)
In our laboratory, both time resolved and steady state luminescence techniques are used. The time resolved technique uses a room temperature germanium photodiode, and the steady state technique uses a liquid nitrogen cooled germanium photodiode.
Time resolved near-infrared luminescence was measured with equipment similar to that described by other workers and is shown in figure 2.7 (Keene et al, 1986; Rodgers and Snowden, 1982).
The detection system comprises a germanium photodiode (EOSS G-O50), with a 5mm2 active area, held near to a 10mm fluorescence cuvette. Selection of the 1270nm luminescence from other infrared emissions was by an interference filter with 82%T at 1270 nm and <0.1%T below 1100 nm. The detector output was amplified using a Judson amplifier, with variable load resistor (0.1-10K½, giving time resolution of ~0.3-2µs for 0.1-1k½), collected on a Tektronix 2432A digital oscilloscope and transferred to a personal computer for data analysis. The solutions of sensitisers were excited at 90_ to the detector by pulsed laser excitation. A BG38 filter was used to remove the 1064nm fundamental laser emission that would otherwise increase the intensity of the fluorescence spike on the transient (figure 2.8).
The decay kinetics and the quantum yield of singlet oxygen generation may be determined by analysis of the transient providing the detector response is sufficiently fast and if the rate of decay of singlet oxygen is significantly slower than its rate of formation. The intensity of singlet oxygen luminescence at time, t, after the laser excitation pulse is given by equation 2.41 (Rodgers, 1988).
The intensity of luminescence at time zero, I, may be extracted by fitting the luminescence decay with a single exponential function back to t=0.
The quantum yields of singlet oxygen generation (F) may be found by plotting Ias a function of laser energy for different sensitisers in the same solvent to give a linear plot with a gradient proportional to the absolute quantum yield (FÆ). The absolute quantum yield may be determined by comparison of the gradient obtained for the sample, with that obtained for a known reference under conditions where laser excitation wavelength, sample absorption, solvent, and detection system arrangement are identical. In quantum yield determinations the sample and reference solutions are optically matched (same absorbance) at the excitation wavelength in order that each solution absorbs the same number of photons, and are excited with an identical number of photons (same laser energy).
For most experiments haematoporphyrin (porphyrin products) (FÆ=0.53, Redmond et al, 1987) was used as a reference in air saturated MeOD, meso(m-sulphonato-phenyl)porphyrin (porphyrin products) (FÆ=0.62, Verlhac et al, 1984) in D2O and meso tetraphenylporphine (porphyrin products) (FÆ=0.58, Bonnett et al, 1988; Rossbroich et al, 1985) or benzophenone (Koch-Light) (FÆ= 0.30, Gorman and Rodgers, 1986) in air saturated benzene.
Deactivation of singlet oxygen by solvent molecules was briefly discussed in section 1.5.5. However, singlet oxygen may also be quenched by an added substrate (a quencher) such as sodium azide(NaN3) in aqueous solutions, diazabicyclooctane (DABCO) in organic solvents, all-trans b-carotene, or even the sensitiser itself.
The effect of an added substrate depends on its concentration and its rate of reaction (by both physical and chemical deactivation) with singlet oxygen. A Stern-Volmer plot of observed rate versus quencher concentration of decay gives from the gradient, the value of kq for the quencher and the value of kd=kr+knr from the y-axis intercept (see equation 2.37).
A recent development in our laboratory has been the introduction of a steady-state near infrared luminescence spectrophotometer (figure 2.9). This is a high sensitivity instrument capable of detecting singlet oxygen generation from sensitisers with quantum yields as low as ~4x10-7 (Fp ~2x10-8). However, this high sensitivity can only be realised in weakly deactivating solvents such as CCl4 and fluorocarbons in which singlet oxygen lifetimes are large. A second drawback of the high sensitivity is that the detector has a slow response time and is thus unsuitable for time resolved work.
The excitation sources used for this method are either a 100W water cooled mercury arc lamp, an argon ion + dye laser system, or a low power helium neon laser.
The 100W water cooled mercury arc lamp (Hanovia) was used with 321, 365 and 405nm narrow band interference filters to isolate the three principle lines at 313, 365, and 405nm. The convergent beam from the arc lamp was focused through a heat filter (either a KG3 heat filter or a 10cm quartz water filter) and collimated to ~1cm at the cuvette.
The argon ion laser (Cambridge Lasers CL5, 5W all lines) was operated in constant light mode for stability (other mode is constant current in which the laser power varies with time but current stays constant), and used with either, a prism attachment for selecting individual argon ion lines (e.g. 480 and 514nm); or a 100% reflecting mirror giving 5W all lines for pumping a dye laser. The dye laser used (Spectra-Physics 375B) was equipped with optics to cover the range 610-910nm (for DCM, LDS751 and LDS820 dyes) although the principle line used was ~650nm. Emission from the laser was attenuated by neutral density filters and focused into an SMA 906 terminated optical fibre (125µM core diameter), with a Newport fibre optic coupler.
The helium neon laser (Spectra Physics, 30mW) was mounted in place of the mercury arc lamp and chopped prior to the 1mm beam being expanded to 10mm with a lens arrangement.
The excitation beam from the arc lamp, or the laser was modulated with a variable frequency optical chopper (Scitec Instruments Ltd, model 300). The square-wave reference signal from the built-in near IR-LED/phototransistor pair was fed to the lock-in amplifier (Scitec Instruments Ltd, model 500MC ) reference signal port.
The emission radiating from the cuvette was collected with a concave mirror (ƒ=50mm) and focused through a hole in the detector housing behind which was mounted a 93% transmission 1270nm interference filter. The 1270nm emission was then collimated by a 13mm focal length lens to fill the 25mm2 active area of the liquid nitrogen cooled germanium detector (North Coast Optical Systems model EO-817L, operating at a bias voltage of -250V). The detector was mounted vertically in a light tight box fitted with a shutter. The output from the integral preamplifier of the detector was also fed to the lock-in amplifier that was interfaced to a personal computer for data collection and storage. The 1270nm luminescence signals were obtained by manually fitting a line through plots of the lock-in amplifier voltage against time, and correcting for the number of photons absorbed at the beginning and end of a scan (usually 5-15 minutes and an average determined). For outputs with good signal to noise ratios, the voltages were simply read off the lock-in amplifier digital display.
For the phthalocyanine phosphorescence emission work, the apparatus was fitted with a 'high intensity' monochromator (~ƒ=7, Bausch and Lomb 33-86-03, blazed at 1000nm, 675 grooves per mm). The monochromator was fitted with a BBC micro computer controlled stepper motor (0.234nm/step). The monochromator was roughly calibrated with 1st and 2nd orders of the helium neon laser using a back slit of 0.3mm giving a band width of <2nm.
The singlet oxygen luminescence quantum yields determined by the steady state method are a product of singlet oxygen quantum yield (FÆ), and the unimolecular decay rate constant (kd). The value of kd becomes more significant in solvents with weakly deactivating properties (e.g. CCl4, and fluorocarbons) because minute impurities can significantly change the value of kd, and hence the value of FÆ. For this reason solvent choice was generally restricted to D2O, MeOD, EtOD, C6H6, and C6D6. Occasionally CCl4 was used when absolute quantum yields were not being measured, but maximum signal intensity was required (e.g. for alignment).
The quantum yield of singlet oxygen luminescence (FL) is defined by equation 2.38 (Gorman et al, 1991; Ogilby and Foote, 1983; Schmidt and Afshari, 1990; Schmidt et al, 1989; Scurlock and Ogilby, 1987).
Where: kd is the rate constant for O2(1Æg) deactivation in solution (kd=kr+knr).
FD is the O2(1Æg) quantum yield.
The intensity of the O2(1Æg) emission, IÆ, in steady state measurements is directly related to FD, kr and tD according to equation 2.39.
Where: A' is an instrumental constant.
tÆ is the lifetime of singlet oxygen.
This method was used to determine singlet oxygen yields for bilirubin, a ruthenium phthalocyanine with a near zero singlet oxygen yield in water, and a chemotherapeutic agent, mitoxantrone, all of which could not be characterised by time resolved techniques. The singlet oxygen yields were determined for this method by fitting a straight line through plots of the lock-in amplifier voltage against time, for air saturated, and for argon saturated solutions of sample and reference compounds (Sodium azide or diaza bicyclooctane were often used as an alternative to nitrogen), the difference in voltage between the two, being attributed to the intensity of singlet oxygen luminescence. These relative intensities were then corrected for slight differences in absorbance at the excitation wavelength and the singlet oxygen yield calculated in the same way as for the time resolved luminescence method.
In addition to the problems involved in the determination of O2(1Æg) quantum yields due to the differing lifetime of singlet oxygen in a particular solvent as a result of quenching, the fluorescence tail can cause a significant error if not corrected for. Although the fluorescence may have a wavelength maximum of about 600-800nm for most porphyrins and phthalocyanines, the tail is still measurable in the wavelength region of the O2(1Æg) emission due to the high sensitivity of the detector. Indeed, in solvents such as water in which singlet oxygen has a low phosphorescence yield this tail may obscure the singlet oxygen emission completely. This problem can however, be overcome in two ways. The first, as described previously is to exchange the oxygen in the solution for argon, or to add a O2(1Æg) quencher such as DABCO or sodium azide. The remaining signal should be due to the fluorescence tail. One problem with using a gas to remove the oxygen instead of chemically quenching the singlet oxygen as it is produced, is that because oxygen is no longer present to quench the triplet states, phosphorescence emission may be observed from the sensitiser giving an incorrect value for IÆ. Furthermore there is no longer oxygen quenching of the singlet state and so fluorescence will also be more intense. The second method of overcoming the problems associated with the fluorescence tail involves adding a 'time resolution' to the steady state measurement. As phosphorescence and fluorescence have considerably different lifetimes, the phase modulation method may be employed through the optical chopper and lock-in amplifier. This is achieved by setting the chopper at a frequency whereby the 'in-phase' signal contains phosphorescence and fluorescence, and the 90_ 'out of phase' signal contains only phosphorescence (assuming no delayed fluorescence), thus separating the two signals (Parker, 1987; Patterson et al, 1990).
The measurement of singlet state lifetimes can add to the characterisation of the photoprocesses of a molecule by providing information such as polarity and viscosity of the environment, together with the nature of quenching processes (Brookfield, 1985; Cubeddu et al, 1989; Sparrow et al, 1986; Yamashita et al, 1984).
Time resolved fluorescence measurements were made on several compounds at the Daresbury laboratory, Station 12.1 (time resolved spectroscopy) with the assistance of Drs. A. N. Macpherson, David Shaw and Moira Behan-Martin, using a technique known as time correlated single photon counting. The excitation source for this work was provided by the synchrotron radiation source (SRS) operated in single bunch mode. Whilst the time resolution (180ps FWHM) and repetition rate (3.1 MHz) does not match some of the fastest laser systems available, synchrotron radiation has the distinct advantage that it is polychromatic. The experimental arrangement for station 12.1 is given in figure 2.10.
Radiation from the storage ring was selected with a computer controlled SPEX1500SP monochromator, adjusted to give an optimum count rate. The fluorescence emission was detected at 90_ to the excitation by a red sensitive photomultiplier cooled to -30_c (through a polariser set at 57.4_, the magic angle, to avoid rotational reorientation effects), the wavelength of interest being selected with bandpass interference filters (FWHM Å10nm). The amplified signal from the photomultiplier was electronically separated from noise, fed to a multichannel analyser, and temporarily stored on a PDP-II minicomputer before being transferred to a NAS-7000 mainframe for permanent storage and analysis. Emission was recorded, where possible for a total of 20, 000 counts and compared to instrumental response obtained by setting the excitation wavelength to the emission wavelength and counting the scatter from a suspension of LUDOX in the cuvette holder.
The fluorescence decays were analysed by a computer program (FLUOR) which employs a weighted least squares curve fitting method (O'Connor and Phillips, 1984). Using this program decay profile is calculated by convolution of the instrument response function with the sum of exponential decays given by,
Where: I is the fluorescence intensity.
ai is the amplitude of a component of
lifetime ti.
The calculated decay profile is then compared to the experimental data, Chi-squared (c2) and residuals indicating the quality of the fits & hence the lifetimes obtained.
Fluorescence emission spectra were measured using a Perkin-Elmer LS50 PC-controlled fluorimeter. Unless otherwise stated the fluorescence spectra were uncorrected for photomultiplier response. Fluorescence quantum yields for air saturated solutions were measured at slit widths of 5nm for both excitation and emission monochromators, the sample always had an absorbance at the excitation wavelength of less than 0.1 to minimise inner filter effects. Except where indicated, quantum yields were measured against haematoporphyrin in methanol (0.09)(Smith, 1985) by comparison of baseline corrected areas under the fluorescence traces.
Where A is the area under the
fluorescence trace,
n is the refractive index,
A is the absorbance at the excitation wavelength.
Last modified on: Thursday, October 30, 1997.