LIGHT AND HIGH ENERGY INDUCED ENERGY TRANSFER IN LIQUID AND RIGID ORGANIC SCINTILLATORS* BY FELIX H. BROWN,? MILTON FURST,$ AND HARTMUT KALLMANN~ Received 5th February, 1959 In contrast to excitation by high energy where the primary energy is absorbed mainly in the bulk material, under light excitation the component which absorbs the incoming radiation may be varied in many solutions by using appropriate wavelengths. Thus the solute and solvent may each be separately excited. By comparing fluorescence under both conditions an absolute measurement of energy transfer occurring from solvent to solute can be made. The results of such determinations show that the probability of energy transfer from effective solvents approaches unity if the solute concentration is sufficient. This occurs particularly when the lowest excitation level of the solvent is energized, but there are indications that the quantum efficiency is near unity when higher levels are excited.Three modes of energy transfer are discussed. Experiments using rigid media aid in discriminating among these since material diffusion transfer does not occur. Definite differences are found in rigid and non-rigid media. Quenching materials are found to be less effective in rigid media. The use of an intermediate " solvent " produces fluor- escence enhancement in rigid as well as non-rigid media. Polymethylmethacrylate and polystyrene show somewhat different energy-transfer properties. It appears that migration transfer plays an important role, at least in rigid media. Transfer of excitation energy from bulk material to solutes has been investigated for many years. This investigation has recently gained impetus from scintillation work ; 1 in this case the bulk material is excited by high-energy radiation in which not only the lowest excited state but also higher states and ionization are produced. These experiments showed that considerable amounts (- 10 %) of the total energy absorbed in the bulk material is emitted by the solute.2 There were also indications that most of the transferred excitation energy comes from the lowest excited state of the solvent molecule.These indications were confirmed by light-excitation experiments which showed that the contributions to transfer made by ionization and higher excitation-energy states are small.In these experiments, it is the solvent which is excited by light ; a portion of the energy absorbed by the solvent is emitted as fluorescent radiation by the solute. Ionization and higher energy states can be completely excluded by choice of suitable wavelengths. When this is done energy transfer is found to proceed to the same extent as under high energy. The fluorescence described is molecular fluorescence and not technical fluorescence which is sometimes de- scribed in the literature.3 This signifies that absorption-reemission effects are negligible in these experiments. From these and other results it can be concluded that the following three modes are most probably responsible for the total energy transfer process from the solvent to the solute. (i) An excited solvent molecule in its movement (diffusion) through the solvent comes close enough to a solute molecule for excitation energy to go over to the * Work supported by U.S. Army Signal Corps Engineering Laboratories, Fort Mon- mouth, New Jersey.j- Dept. of Physics, New York University, New York 3, New York. 2 Dept. of Physics, New York University, New York 3, New York, and Dept. of Physics, Hunter College, Bronx, New York. 4344 ENERGY TRANSFER I N ORGANIC SCINTILLATORS solute molecule during the period of time they are near each other. This mode is called material diffusion transfer. (ii) The excitation energy (rather than the molecule) migrates through the solvent by jumping from one solvent molecule to a neighbouring one. Eventually the excitation energy comes close enough to a solute molecule, and the energy goes over to the solute.This type of energy transfer is designated migration transfer. (iii) There may be neither diffusion nor migration of excitation energy in the solvent, but instead the energy of the solvent molecule may be transferred directly to the solute in a single step (jump) over relatively large distances. This mode is designated single-step transfer.4 All three processes may occur in the solution with one being perhaps of major importance in a given solution. Another type of energy transfer, namely that by absorption of light and re- emission has been proposed.5 In the cases we have investigated there is practically no energy transfer by fluorescent radiation ; only the three processes mentioned above occur.LIGHT EXCITATION OF LIQUID SYSTEMS The use of light excitation in measuring energy transfer has limitations not present when high energy is used. In dilute solutions under high energy excita- tion, the primary energy is absorbed mainly in the solvent because its mass is so much greater than that of the solute. When light excitation is used the absorption spectra of solvent and solute must be suitably matched for this to occur. For occurrence of considerable energy transfer the solute concentration must be of the order of 0.01 M ; this is known from high energy measurements.6 Con- sequently at such concentrations the ratio of absorption of solute to solvent must be about 0.02 for practically all the energy to be absorbed by the solvent. Another limitation is that a number of solvents, worthy of study from the viewpoint of energy transfer, cannot be as easily excited by light because their absorptions lie too far in the ultra-violet. A notable example is p-dioxane, interesting because it has good energy transfer properties under high energy despite its lack of a con- jugated double-bond system, generally found in solvents which transfer well.After considerable experimentation, it was found that an adaptation of the Beckman DU spectrophotometer makes possible reliable measurements of energy transfer under light excitation. The basic experiment was a comparison of the fluorescence of the solute directly excited by the incoming radiation with its fluorescence when excited by energy transfer from the solvent.The direct light was essentially completely absorbed in the soIution in most cases. Possible stray light effects were checked by measuring energy transfer in the same solutions by reflection as well as transmission. On the whole the results were the same. These reflectance measurements were also made in order to decide the possible influences of geometry effects. In the transmission measurements the light was incident normal to the solution. The photomultiplier was along the line of incidence. A filter which did not pass the direct incident light was placed in front of the photomultiplier, which filter passed the fluorescence radiation ; in any event the direct light is almost entirely absorbed in the solution. In the re- flection arrangement, the photomultiplier and the incident light were on the same side of the solution, and the light did not go through any vessel.The overall results of these measurements were quite similar for " effective " and " poor " solvents.7 Both were essentially independent of exciting wavelength, and the relative intensities were the same at different concentrations. Since it is more convenient to use the transmission arrangement, in part because of greater in- tensity, most of the light excitation measurements were made in this way. Quartz cells with a 1 cm path length were used and the light source was a hydrogen lamp. The fluorescence of a given solute concentration was measuredF . H . B R O W N , M . FURST A N D H . KALLMANN 45 in a solvent not absorbing the incoming radiation and in another solvent which absorbs the incoming radiation. The determination of solute fluorescence in the non-absorbing solvent serves as a measure of the intensity of the incident light.(In some cases the fluorescence efficiency of the solute is not the same in both solvents as determined by comparing the fluorescent light output of the Same solute in both solvents using exciting wavelengths not absorbed by either solvent.) If the ratio of the light outputs of both solutions at a given wavelength is the same as when no light is absorbed by the solvent then there is 100 % energy transfer ; there is then no decrease of fluorescent light when the light is first ab- sorbed in the solvent and thereafter transferred to the solute. In fig. 1, 2 and 3, the ratio of the fluorescent light outputs is given for all wavelengths used ; those at which neither solvent absorbs are always to the right.exciting wavelength (A) FIG. 1.-Fluorescence of PBD in anisole and in pdioxane under u.-v. excitation. As an example, at X = 2500 A, in cyclohexane + fluoranthene (1 g/l.), cyclo- hexane does not absorb the incoming radiation whereas fluoranthene does. On the other hand, in anisole + fluoranthene (1 g/l.) the anisole absorbs more than 99 % of the incoming radiation. Even at higher fluoranthene concentrations little of the incident light is absorbed by the solute. In both solvents the fluor- escence emitted is that characteristic of fluoranthene. In cyclohexane there is no energy transfer from the solvent when the exciting wavelength is 2500A.In the other solvent (anisole), fluorescence occurs via energy transfer at excitation wavelengths less than 2800A since these wavelengths are absorbed only by the solvent. The ratio of the solute light output under energy transfer conditions to that when direct excitation occurs is a measure of energy transfer. Measurements show that the energy transfer in some solutions is very efficient and approaches 100 %. At large solute concentration some fraction of the incident light is also absorbed by the solute (e.g. lOg/l. of fluoranthene in fig. 2 at 2500A). In order to determine the exact amount of energy transfer, a cor- rection must be made for this direct absorption. The correction was not made in these figures since it is not large except in the wavelength region near the absorption edge of the solvent.Light-induced transfer was measured at different exciting wavelengths. This was done in order to determine the effect of higher solvent excitation levels. Pre- liminary work seemed to indicate that higher levels were less efficient than lower ones.4 These early measurements were not so accurate as those herein reported especially near the surfaces of the solution. The present, more reliable nieasure- ments indicate, however, that energy transfer is essentially independent of exciting46 ENERGY TRANSFER IN ORGANIC SCINTILLATORS wavelength. While wavelengths obtainable from the hydrogen lamp do not allow great variations in the electronic excitation levels, considerable differences in vibrational levels certainly occur. In anisole an electronic excitation level above the first excited level is also reached.Fig. 1-3 describe energy-transfer measurements under light excitation of wave- lengths down to 2300A. Fig. 1 and 2 utilize a solvent (anisole) known from measurements under high energy excitation to be effective for energy transfer. exciting wavelength (A) FIG. 2.-Fluorescence of fluoranthene in anisole and in cyclohexane under u.-v. excitation. I I I I 1 I I I I I I 7 4 0 0 2500 2600 2 7 0 0 2 8 0 0 2 9 0 0 3 0 0 0 3 1 0 0 3200 3300 3400 exciting wavelength (A) FIG. 3.-Fluorescence of fluoranthene in anethole and in cyclohexane under u.-v. excitation. Fig. 1 depicts results with 2-phenyl-5-(4-biphenylyl) oxadiazole [PBD]. This solute is the most efficient known; it emits in the near ultra-violet.Fluoranthene, shown in fig. 2, is a solute of medium efficiency, fluorescing in a wavelength region extending into the visible. The curve of fig. 2 at 10 g/l. of fluoranthene is almost horizontal and the same for wavelengths absorbed by the solute or by the solvent. Practically 100 % transfer occurs for most of the wavelengths below 2800& since the solvent absorbs nearly all of the incoming radiation. Fig. 3 describes experiments in a solvent (anethole) known from high-energy measurements to be " moderate " in its transfer properties. It is seen that under light excitationF . H . BROWN, M. FURST A N D H . KALLMANN 47 the energy transfer from the “moderate” solvent is also smaller than from the “ effective ” one. Even with 20 g/l.solute the fluorescence is considerably smaller when the energy is absorbed by the solvent than when it is absorbed by the solute. It is to be noted that anethole quenches the solute to a sizeable extent SO that the ratio at the right side is not unity. Some energy is absorbed by the solute at the highest concentration, so that the energy transfer is smaller than indicated by fig. 3. Similar results were obtained using 1-methylnaphthalene and p-xylene as solvents. Up to the present, few compounds have been tried partially because of the limitations discussed above. I I I I I ..A 5 10 15 I W PBD conc. (g/1.) FIG. 4.-Fluorescence of PBD in p-xylene under u.-v. and gamma-ray excitations. A direct comparison between energy transfer produced by light excitation and by high energy excitation has also been made.The exciting light source was alternated with a high energy exciting source, and the rest of the measuring ap- paratus was left unchanged. The results are shown in fig. 4 for y-radiation, for 2400 A, 2700 A and 3100 A exciting wavelengths. Energy transfer occurs in all cases except that of 3 100 A excitation because this light is not absorbed by the solvent. Excluding the last, the three curves coincide almost throughout ; only in the middle range are there small deviations, probably within experimental accuracy. The curves indicate that energy transfer produced under high energy is identical to that produced under light excitation, and that in both cases it occurs from the lowest excited state of the solvent.These experiments were undertaken to investigate whether the previously ob- served difference between gamma ray and light induced fluorescence was real or due to different arrangements. In the present measurements, geometry was the same in all cases, and it was ascertained that the results were independent of48 ENERGY TRANSFER I N ORGANIC SCINTILLATORS geometry. Thus we conclude that there is very little difference in energy transfer produced by gamma ray and light excitation and that the previously found differ- ence was due to geometric effects. Energy transfer occurs via the lowest excited state of the solvent; it is almost 100 % when light excites the solvent at high enough solute concentrations. This means that for 100 quanta absorbed in the solvent, 100 are transferred to the solute.Energy transfer produced under high energy and light excitations of various wavelengths are practically identical. This excludes considerable contribution of higher states and of ionization to energy transfer. The results also seem to show that there is 100 % transition from the second electronic level to the first, since under light excitation with 23008, the second electronic level of anisole is excited and no difference in transfer is observed (see fig. 1 and 2). The next problem is which of the three processes described in the introduction is most important for energy transfer. Experiments in rigid media were per- formed in order to find out something more about these processes. They are described in the following sections.The following conclusions can be drawn from these experiments. INVESTIGATIONS IN RIGID MEDIA It is possible to discriminate between the three processes as follows : process one, transfer by material diffusion, is excluded in rigid media ; thus only processes two and three can occur. Processes enumerated for energy transfer are also re- sponsible for quenching, in which the excited state of the solvent or solute is non-radiatively transferred to the ground state by interaction with some other molecule. The excited solvent molecule in a rigid medium can be quenched according to process two. The excitation energy migrates from solvent molecule to solvent molecule and eventually reaches the neighbourhood of a quencher where it undergoes a radiationless transition to the ground state.When migration of energy does not occur, as is the case for an excited solute, then energy quenching can occur only via process three. Since quenching is often not connected with a resonance process, it occurs when the excited solute molecule happens to be in the vicinity of a quenching molecule. Energy migra- tion from solute to solute molecule does not take place because of the large distance between these molecules, and transfer via radiation is excluded because little or none of its own radiation is absorbed by the solute. (Absorption effects can be determined by varying the sample thickness.) Thus for solute quenching in rigid media, only process three is significant, and only those excited so'ute molecules are quenched which are near a quenching molecule.The formula for solute quenching, occurring according to process three is (1) N is the number of molecules quenched per second ; c is the solute concentration, P gives the number of solute molecules excited per second, and VO(= $ 7 4 ) is the so-called quenching volume (assuming as a first approximation that when the distance between excited solute molecule and quencher is greater than YO, no quenching occurs, and when it is less than ro quenching occurs). Light emission by the unquenched molecules is then proportional to P exp (- voc). Energy transfer according to process three can be described by a formula like (1) in which c becomes the concentration of the substance to which energy is transferred, and vo the transfer volume, which is generally larger than the quenching volume which can be assumed to be of the order of the molecular radius cubed.This implies that quenching occurs only when a quenching molecule is within about 5 x 10-8 cm from the centre of the excited molecule. The eth part of the molecules are quenched at a quencher concentration c = l/q, generally of the order of 1 mole/l. In liquid media, quenching occurs at concentrations of about 0-1 M or even smaller but in these media quenching also occurs via process one N = P[1 - exp (- voc)].F . H . BROWN, M. FURST A N D H . KALLMANN 49 and from this process a larger quenching probability results than from process three. The value of v depends upon the lifetime of the excited molecule in all cases-only slightly in a non-resonance process and much more in a resonance process.Experimentally it is found in accordance with the above that solute quenching in rigid media is much smaller than in liquids (see table 1). TABLE SOL SOLUTE (0,045 M PPO) QUENCHING BY ~,CY.,X,S(’,X’,CC’-HEXACHLORO-~-XYLENE IN XYLENE AND POLYSTYRENE remaining solute fluorescence as percent of unquenched solute fluorescence xylene, yo polystyrene, yo quencher concentration 0.016 M 48 96 0.032 M 36 94 0-064 M 29 89 There may be a considerable difference between quenching of the solute and of the excited solvent molecule in a rigid medium since in the latter case quenching also can occur by energy migration (process two). Experiments indeed indicate that the solvent is more quenched than the solute. The next step is the investigation of energy transfer itself.In rigid media it is only necessary to discriminate between processes two and three. Energy transfer according to process two in liquids is given 9 by eqn. (2) (which also would hold for process one with a different value of Q) in which R + c in the denominator describes the concentration quenching of the solute. This term is unnecessary in rigid media because of the absence of concentration quenching which requires process one (see below). Pc P’c <Q+c)(R+c) --+ ’ I = I = P*[l - exp (- vc)]. (3) Eqn. (3) gives the energy transfer according to process three (same form as eqn. (1) discussed above). In this case, Y is the distance over which energy can be transferred during the lifetime of the excited solvent molecule. This distance can be quite large so that process three may contribute considerably to energy transfer even in rigid media.The value of r for transfer is greater than that of YO for solute quenching which latter is of molecular dimensions. Thus, according to formula (3), energy transfer may occur at relatively small concentrations, whereas solute quenching according to formula (l), occurs only at larger concentrations. The experiments were per- formed with polystyrene (PS) and polymethylmethacrylate (PMMA) as solvents. The results obtained from these polymers are separately discussed. Energy jumps through distance Y in one step. POLYSTYRENE (Ps) Quenching experiments yielded the following. Concentration quenching in polystyrene does not occur at any measured concentration.It also does not occur when large amounts of an intermediate “ solvent ” such as naphthalene are added. This bears out the contention that concentration quenching is due to material diffusion (process one). The absence of concentration quenching is evident in fig. 5, which depicts high energy induced fluorescence of 2 : 5-diphenyl- oxazole (PPO) in polystyre -e and in polystyrene plus 0.2 M naphthalene as inter- mediate solvent. There is no decrease of fluorescence at higher concentrations ; such a decrease is observed in liquids. Fluorescence under light excitation of the solute has also been measured, and the absence of concentration quenching confirmed.* * Such quenching occurs with triplet states because of their longer lifetimes.50 ENERGY TRANSFER I N ORGANIC SCINTILLATORS The next step was to introduce quenchers into polystyrene systems in order to measure their effect on the fluorescence under gamma rays and under direct solute excitation by light.These experiments show clearly that solute quenching by a PPO conc. (mole/]. monomer) (0-2 M)/PPO. FIG. 5.-Gamma-ray induced fluorescence of 0 PS/PPO and PS/naphthalene quencher conc. FIG. 6.-Gamma-ray induced fluorescence of PPO scintillators quenched by or., ay a, a‘, a’, a‘-hexachloro-p-xylene. 1. PS/naphthalene (0.2 M)/PPO (0.045 M). 2. PS/PPO (0.045 M). 3. XylenelPPO (0.023 MI. given compound is smaller in polystyrene than in liquid solvents. This is con- sistent with the absence of processes one and two for solutes in polystyrene solu- tions. Diphenylmercury and m,cc,a,m~,a’,cc’-hexachloro-p-xylene were used as quenchers.(Fig. 6 depicts results with the latter.) Both substances when tested in liquid media were found to be strong quenchers for the solute, 2 : 5-diphenyl-F. H . BROWN, M . FURST AND H . KALLMANN 51 oxazole (PPO). In polystyrene, however, a,a,a,a’,a’,a‘-hexachloro-p-xylene pro- duced almost no solute quenching and diphenylmercury quenched this solute by only about 20 oh at the relatively high concentrations employed. This indicates that the quenching radius of diphenylmercury in polystyrene is of the order of 10-7 cm according to formula (1). Solvent quenching is discussed later since it is related to the mechanism of energy transfer. Two basic types of energy-transfer experiments were performed.In one, polystyrene containing various solutes was excited by gamma rays, and the fluorescence determined as function of solute concentration ; in the other, similar solutions were measured with various amounts of naphthalene present as inter- mediate solvent (fig. 5). In these latter experiments the excitation energy of polystyrene is assumed to be transferred to naphthalene and then from naphthalene to the solute, similar to the sequence found in analogous liquid solutions. This assumption is borne out by experimental results (see below). Experiments without naphthalene show that polystyrene transfers energy to the solute quite well but larger solute concentrations than in “ effective ” liquid solvents are required. The total amount of energy transferred is a little smaller than in liquids.This was determined by comparing the light efficiency of the same solute in xylene and polystyrene under light excitation and high energy excitation. It may be recalled that the concentration at which energy is transferred is essentially determined by the speed of energy transfer and the lifetime of the excited solvent molecule. Since it is known from decay time measurements that this Metime is even longer in polystyrene than in xylene,lo one concludes that the necessity for larger solute concentration in polystyrene is due to an impairment of the energy transfer process itself. The simplest assumption would be that material diffusion, process one, does not occur in polystyrene, but is responsible for part of the energy transfer in xylene, and that energy migration, process two, in polystyrene is not so effective as processes one and two together in xylene.This assumption is not sufficient to explain all the results since it is found that the shape of the fluorescence against concentration curves observed in polystyrene containing effective solutes, with the exception of some special cases discussed below, is different from those of a liquid solution.11 They cannot be completely explained by either process two or three. It may be that a combination of both processes is responsible for the behaviour. Addition of naphthalene to polystyrene solutions has two effects. Smaller solute concentrations are required for maximum energy transfer. (0.2 M naphthalene is sufficient to bring about maximum effect.) Secondly, the shape of the concentration curve is altered.The experimental results in polystyrene solutions containing added naphthalene can be better represented by formula (3) (process three). The maximum light emission is the same with and without naphthalene. Smaller solute concentrations required for maximum fluorescence may be due to a longer lifetime of excitation energy in naphthalene compared to polystyrene. It is known that naphthalene and its derivatives in liquid solu- tions have somewhat longer lifetimes than does polystyrene.10 Lifetime differ- ences may, of course, not be the only factor. There is, however, another noteworthy point. It is observed, as mentioned above, that the change in solute concentration required for maximum fluorescence caused by addition of naphthalene is terminated at relatively small naphthalene concentrations (0.2 M).More naphthalene does not change the concentration curves further. At this naphthalene concentration all possible excitation energy is transferred from polystyrene to naphthalene. This is borne out by the ob- servation that, at this same naphthalene concentration, maximum fluorescence from naphthalene occurs under high-energy excitation in the absence of other compounds.11 If it is assumed that energy transfer from naphthalene to naph- thalene occurs, it can only occur via migration transfer (process two). Then the52 ENERGY TRANSFER I N ORGANIC SCINTILLATORS following difficulty arises. It would be expected that migration from naphthalene to naphthalene be bettered by continued increases in naphthalene concentration ; this would decrease the solute concentration required for a given amount of energy transfer.The absence of this decrease is a prominent feature of all solutes investigated in polystyrene plus naphthalene. The process of excitation energy migration from naphthalene to naphthalene when the molecules are not so close to each other as they are in solid or liquid naphthalene is not a very probable one although it is a resonance process. The reason is that the transition from the excited state (no vibrational excitation) to the ground state usually includes excitation of vibrational levels of the latter state. A transition to a vibrational level of the ground state does not release enough energy to excite another naphthalene molecule.Since the transition to the ground state without excitation of vibrational levels includes only a small portion of the total transition probability, the probability for a migration process entirely based on dipole transitions in naphthalene molecules is not very large. This probability of migration is further reduced because the total dipole transition moment for the lowest excited state of naphthalene is extremely small. A rough evaluation based on these considerations shows that energy should not migrate from naphthalene to naphthalene in a rigid medium over distances of 2 x 10-7 cm (equivalent to 0.2 M) during the lifetime of the excited naphthalene molecule. This leads to the conclusion that transfer from naphthalene to solute does not occur by energy migration (naphthalene -+ naphthalene + solute) but rather by a single-step transfer (process three) ; energy is directly transferred from naph- thalene to solute over relatively large distances.The probability of this process is much larger than that between molecules of naphthalene for two reasons. First, the respective dipole transition probability of the accepting solute molecule is larger than that of naphthalene, and the transfer to a solute is better than to another naphthalene molecule since the transfer probability is a product of two transition probabilities. Secondly, the difference in energy levels between naphthalene and solute is important. Because of this difference not only are zero-zero transitions in naphthalene capable of producing transfer to the solute, but also transitions to higher vibrational levels of the naphthalene ground state.Thus it is understandable that energy transfer from naphthalene to solute in rigid media does not occur via migration transfer but via single step transfer. These ideas follow Foerster’s.4 One would be inclined to assume a similar process for transfer of energy in polystyrene without naphthalene. But as already stated this is not borne out by the fluorescence against concentration curves observed in polystyrene. The only explanation we can offer at present is that processes two and three both occur. This will be investigated in more detail, especially by measuring lifetimes. We hope to report on this in the near future. In liquid systems the transfer parameter Q displays an amazingly small de- pendence upon the solute and its transition probability to the ground state as determined from absorption spectra.This may be due to the predominance of processes one and two, in which the ultimate transfer may occur between molecules very near each other. In polystyrene + naphthalene systems, however, in which it is assumed that process three predominates in the final transfer to the solute, it would be expected that v is more strongly dependent upon the solute used. More information about these processes is derived from experiments on the quenching of solvent molecules. Such quenching can be studied by observing the quenching of high energy induced fluorescence. It can be concluded that most of the fluorescence decrease observed is due to solvent quenching since it was determined that the solute is only relatively slightly quenched.When di- phenylmercury or ct,qx,a’,ct’,ct’-hexachloro-p-xylene is added to various poly- styrene systems, a considerable decrease in fluorescence is observed. This quenching is so strong that it is doubtful whether process three alone can accountF . H. BROWN, M. FURST AND H . KALLMANN 53 for it ; the quenching cross-section of these compounds for polystyrene would have to be unusually large. This strong quenching of polystyrene is therefore assumed to be due in part to migration of excitation energy in polystyrene. Here, indeed, migration of excitation energy can occur, since polystyrene molecules are close to each other. Further, transfer of excitation from one polystyrene molecule to another is probably due not only to dipole transitions but also to overlap of wave functions which may be appreciable for neighbouring atoms and molecules.Thus it seems that the strong quenching seen in fig. 6 comes about because excitation energy migrates in polystyrene until it comes close enough to a quencher molecule. The quenching behaviour is different when naphthalene is added. Large concentrations of naphthalene eliminate a sizeable portion of the fluorescence decrease ; naphthalene is considerably less quenched than polystyrene. This is at least partially attributed to the lack of energy migration between naphthalene. The excitation energy has no chance to come near the quencher molecule during its lifetime if the energy initially is localized far from the quencher.The number of naphthalene molecules close to a quencher molecule is not very large at the concentrations employed. Therefore this lack of quenching in naphthalene is a corroboration of the absence of energy migration in naphthalene. The results so far discussed are consistent with our previous ideas of energy transfer.12 There are some results, however, which seem incapable of being explained with these ideas alone. Some solutes have been found to have different energy-transfer behaviour. One such case is fluoranthene in polystyrene without and with added naphthalene. In fig. 7 at small concentrations a rather steep rise of high energy induced fluorescence, and then an almost continuous rise of this fluorescence over a large concentration range is observed.The difference between fluoranthene and other solutes is that energy transfer to fluoranthene increases constantly and does not reach a limiting value at the large concentrations used. If naphthalene is added, the fluorescence against concentration curve is parallel to the curve without naphthalene but the intensities are greater ; this means the same amount of energy transfer is accomplished at smaller solute concentrations. Similar behaviour is shown by chrysene. A simple interpretation would be that energy transfer from polystyrene (no naphthalene) to these solutes is poor, requiring larger solute concentrations than do other solutes. The same must then be true when naphthalene is present since the fluorescence curves are parallel.This is difficult to understand since with considerable amounts of naphthalene added, energy transfer from naphthalene to the solute would not be influenced any more by polystyrene and since in liquid solutions energy transfer to fluoranthene and chrysene is normal. One cannot simply interpret these fluorescence curves as due to a quenching, assuming, for instance, that some of the solute molecules are in positions where they are more easily quenched than in other positions and that this results in a decreased fluorescence. This assumption is untenable since under light excitation the fluorescent behaviour is quite normal; that is, a steep rise of fluorescence in- tensity in the range where absorption becomes completed and then a constant fluorescent light output.Therefore it is not the fluorescence efficiency of the solute which is different, but rather the energy-transfer process. It is noteworthy that the maximum light output of these polystyrene solutions under high energy is higher than that of respective liquid solutions. This higher light output is in accord with the higher fluorescence efficiency of these solutes in polystyrene observed when directly excited by light.11 The only explanation which we can present up to now is that the transfer of energy from the solvent to these solutes over large distances is small because of small transition moment. Calculations based on absorption strengths do not bear this out. Fluoranthene does have a small transition moment (shown by its long time-constant in liquid solutions,10~ despite the fact that fluoranthene is quenched), but this small transition probability is not sufficient to explain the54 ENERGY TRANSFER IN ORGANIC SCINTILLATORS difference between the fluorescence against concentration curves of, for example, fluoranthene and 2 : 5-diphenyloxazole.Furthermore, it would be expected that, in polystyrene, energy transfer occurs partially via migration of energy (process two), but then the smaller transition moment of fluoranthene would be much less important since excited polystyrene molecules can always be close enough to the fluoranthene molecule. In this respect it is also important to consider the curve when naphthalene is present which shows the same behaviour. This indicates that the same impeding effects occur for energy transfer from naphthalene to fluoranthene as from polystyrene to fluoranthene.At this stage it is difficult to pin down what impairs the energy transfer to these compounds in polystyrene. I I I 1 I I 1 I I ' OlOl ' 0 . 0 3 0.05 0 . 0 7 0.09 fluoranthene conc. (moles/l. monomer) FIG. 'I.-Gamma-ray induced fluorescence of PS/fluoranthene and PS/naphthalene + fluoranthene. 1. PS/naphthalene (0.2 M)/fluoranthene. 2. PS/fluoranthene. The results obtained from rigid polystyrene can be summarized as follows. Solute quenching is relatively small and presumably occurs only by process three. Quenching of polystyrene is greater ; it seems to occur by processes two and three. Quenching according to process two does not seem to occur for naphthalene when present as intermediate " solvent ".Even at relatively large naphthalene concentrations, it seems to be small ; so that naphthalene quenching goes on only by process three. Because process one is excluded, solute concentra- tion quenching (and shifts of emission spectra by formation of fluorescent dimers, as is found with pyrene in liquid solutions 13) does not occur. Some solutes in polystyrene, e.g. fluoranthene and chrysene, have a higher fluorescence efficiency when excited by light than they do in liquid solvents. This effect is similar to the viscosity effect which is reported for fluorescence light output of solutes in viscous media,l4 but the effect is not general. Polystyrene energy transfer is different from that from liquid solvents.Not only are larger concentrations required, but the shape of the fluorescence curve is quite different. This is tentatively interpreted as due to a combination of pro- cesses two and three in polystyrene solutions. Energy transfer via an intermediate " solvent " occurs in polystyrene as well as in liquids.F . H . BROWN, M . FURST A N D H . KALLMANN 55 It is conjectured that the transfer from naphthalene to the solute occurs by process three because this transfer reaches its maximum at such low naphthalene concentrations that migration between naphthalene molecules is improbable. POLYMETHYLMETHACRYLATE (PMMA) Experiments similar to those on polystyrene solutions were performed on polymethylmethacrylate (PMMA). Fig. 8 shows the high-energy induced fluor- escence of 2 : 5-diphenyloxazole in PMMA without and with naphthalene as intermediate “solvent”. The overall shapes of the curves are similar to cor- responding ones in polystyrene, but marked differences are present.Because of its structure, it was expected that PMMA be a “ poor ” solvent (with respect to PPO conc. (moles/l. monomer) FIG. 8.-Gamma-ray induced fluorescence of PMMA/PPO and PMMA/naphthalene/PPO. 1. PMMA/naphthalene (0-8 M)/PPO. 2. PMMA/PPO. energy transfer). This would be seen by a very gradual rise with solute con- centration of high energy induced fluorescence. Addition of naphthalene as inter- mediate ‘‘ solvent ” would be expected to make the rise considerably steeper, and relatively large naphthalene concentrations would be necessary to extract all the available energy from PMMA.These ideas were arrived at because PMMA does not contain a conjugated double-bond system (in most cases important to effective energy transfer) and because PMMA by itself exhibits even smaller fluor- escence under high-energy excitation than effective solvents. This latter may be considered as an indication that considerable quenching occurs, making the life- time of the excited state relatively short. The figure shows, on the contrary, that PMMA without naphthalene reaches its maximum fluorescence at relatively small solute concentrations, the same order as those encountered in polystyrene. This would make PMMA a “ moderate ” solvent for energy transfer. The maximum light output obtained from PMMA solutions under these conditions is only about one-third as great as that obtained from polystyrene solutions of the same solutes. The addition of various amounts of naphthalene not only shifts the rise of the fluorescence concentration curve to smaller concentrations as expected, but also raises the maximum light output to about twice that without naphthalene. Larger amounts of naphthalene are needed to extract all available energy from PMMA than are needed in polystyrene; maximum fluorescence is obtained with ap- proximately 1 M naphthalene.A possible interpretation of these results is that two different excited-energy states exist in PMMA with few transitions occurring between them; one having56 ENERGY TRANSFER IN ORGANIC SCINTILLATORS moderate transfer properties, due perhaps to a lifetime of medium duration, the other having poor transfer properties and probably a short lifetime.The excitation energy is extracted from the better transferring state by relatively small concen- trations of solute. Once this energy is extracted, the fluorescence appears to reach a saturation. The second state transfers so poorly that at the available solute concentrations very little of its energy is extracted. This second energy com- ponent, however, can be extracted by large amounts of naphthalene. The need for large naphthalene concentrations is in accord with the assumption that energy in the other excited state is less transferable. When large amounts of naphthalene are present, the PMMA + naphthalene + solute system emits only about 10 to 30 % less energy than the corresponding polystyrene system. The validity of this model for energy transfer from PMMA will be investigated by fluorescence measurements under light excitation of solvent and solute and by concomitant measurements of time constants. 1 Kallmann, Physic. Rev., 1950, 78, 621. Reynolds, Harrison and Salvani, Physic. 2 Broser, Kallmann and Martius, 2. Naturforsch., 1949, 4a, 204. Furst, Kallmann 3 Birks, Physic. Rev., 1954, 94, 1567. 4 Forster, Fluoreszenz Organischer Verbindungen (Vandenhoeck und Ruprecht, 5 Kallmann and Furst, Physic. Rev., 1951, 81, 853. Birks, Scintillation Counters 6 Furst and Kallmann, Physic. Rev., 1952, 85, 816. Furst, Kallmann and Brown, 7 Furst and Kallmann, Physic. Rev., 1955, 97, 583. Furst and Kallmann, J . Clzem. 8 Brown, Furst and Kallmann, J. Chim. Physique, 1958,55, 688. 9 Furst and Kallmann, Physic. Rev., 1952, 85, 816. Rev., 1950, 78,488. Ageno, Chiozotto and Querzoli, Acc. dei Lincei, 1949, 6, 626. and Kramer, Physic. Rev., 1953, 89, 416. Gottingen, 1951). (McGraw-Hill, New York, 1953). J. Chem. Physics, 1957, 26, 1321. Physics, 1955, 23, 607. 1O(a) Kallmann and Brucker, Physic. Rev., 1957, 108, 1122. 11 Brown, Furst and Kallmann, J. Int. Atomic Energy Agency, in press. 12 Kallmann and Furst, in Liquid Scintillation Counting Conference (Pergamon Press, New York, 1958), p. 3. Brown, Furst and Kallmann, J. Chim. Physique, 1958, 55, 688. Kallmann, Furst, and Brown, in Semiconductors and Phosphors (Interscience Publishers, New York, 1958), p. 269. (b) Swank and Buck, Rev. Sci. Instr., 1955, 26, 15. 13 Forster and Kasper, 2. Elektrochem., 1955, 59, 976. 14 Kallmann, Furst and Brown, in Semiconductors and Phosphors (Interscience Pub- lishers, New York, 1958), p. 269.