Synthetic information on pyranylidene compounds 17 (see Fig.6).
Low molecular mass organic compounds with internal charge transfer properties are widely adopted for organic photonics such as materials for the creation of molecular electronics elements, organic magnets, solar cells and organic light emitting diodes (OLEDs) for full display panels [1-3]. One of the most widely used red light-emitting materials contains pyranylidene (4H-pyran-4-ylidene) or isophorene (5,5-dimethylcyclohex-2-enylidene) fragments as backbone of the molecule (see Fig.1), which are conjugated in a system with electron acceptor and electron donor fragments [1,4-24]. In many cases the light-emitting layer from such commercially available compounds is prepared by thermal evaporation in vacuum [1-2, 25-27]. Some of them are used as dopants in a polymer matrix and spin-coated onto a hole transport layer from solution [1,12]. However the doping amount of luminescent compound is limited by self crystallization and photoluminescence quenching at higher concentrations which reduce the quantum efficiency of fabricated devices significantly [11-12]. Therefore it is important to synthesize low molecular mass light-emitting organic compounds which do not crystallize and form thin amorphous solid films from volatile organic solvents. Such compounds, which can make a solid-state glassy structure prepared from solutions, could facilitate technological processes in the production of many devices in optoelectronics, for example, light emitting devices by low-cost deposition such as wet casting methods and easier light-emitting material synthesis. Some of these red light-emitting compounds have been introduced by us [28-32].
In this chapter we present complete synthesis, thermal, optical, photoelectrical and glass forming properties of new organic glass-forming pyranylidene and isophorene fragment containing derivatives with bulky trityloxy groups in their molecules. The optical properties, both in solution and solid state, are compared. The dependance of photoelectrical properties and energy structure of glassy films on molecular structure will be discussed. The most popular derivatives of pyranylidene and isophorene used in OLEDs are shown in Fig.1.
The synthesis procedure of pyranylidene and isophorene D-π-A type luminophores (see Fig.1) with either one or two electron donor fragments can be divided into three main parts:
Synthesis of a backbone fragment: Synthesis of derivatives of 4H-pyran-4-one, which in their molecules contain not only a carbonyl group, but also at least one methyl group and are able to react further with aromatic aldehydes.
Addition of an electron acceptor fragment to the backbone: Condensation reaction of 4H-pyran-4-one derivatives synthesized in 1) with active methylene group containing compounds.
Synthesis of pyranylidene and isophrene D-π-A type red emitters: Final addition of electron donor group containing aromatic aldehydes to compounds obtained in 2).
2.1. Synthesis of the backbone fragment: 2,6-disubstituted-4H-pyran-4-ones
The simplest of 2,6-disubstituted-4H-pyran-4-ones is 2,6-dimethyl-4H-pyran-4-one (compound 2 in Fig.2), which is obtained in 86% yield from dehydroacetic acid (compound 1 in Fig.2) by acidic rearrangement with following decarboxylation (see Fig.2) [32-33].
2,6-Dimethyl-4H-pyran-4-one (compound 2 in Fig.2) has one carbonyl group which can further react with active methylene group containing compounds in Knoevenagel condensation reactions. It also has two activated methyl groups, which can react in the same type of condensation reactions with one or two molecules of aromatic aldehydes.
Another method for the syntheis of 2,6-disubstituted-4H-pyran-4-ones, which contain at least one active methyl group, is using 4-hydroxy-6-methyl-2H-pyran-2-one (compound 3 in Fig.3) as starting material [11,34]. Its further reaction with isobutyryl chloride (compound 4 in Fig.3) in trifluoroacetic acid (TFA) gives 6-methyl-2-oxo-2H-pyran-4-yl isobutyrate (compound 5 in Fig.3). Without separating the compound 5 from the reaction mixture it was subjected to Fries rearrangement resulting in 4-hydroxy-3-isobutyryl-6-methyl-2H-pyran-2-one (compound 6 in Fig.3). In its decarboxylation and further acidic cyclization reactions 2-isopropyl-6-methyl-4H-pyran-4-one (compound 8 in Fig.3) is obtained with 80% yield. Compound 8 also contains a carbonyl group, just as the previously synthesized 2,6-dimethyl-4H-pyran-4-one (compound 2 in Fig.2). Since it now contains just one activated methyl group, only one aromatic aldehyde containing fragment can be added to the backbone of pyranylidene derivative 8 (shown in Fig.3).
One of the most preferred 2,6-disubstituted-4H-pyran-4-ones is 2-tert-butyl-6-methyl-4H-pyran-4-one (compound 13 in Fig.4) [7,11,35]. The first synthesis method starts from 3,3-dimethylbutan-2-one (compound 9 in Fig.4). Treating it with acetic anhydryde (Ac2O) and boron trifluoride (BF3) a boron enolate (compound 10 in Fig.4) is obtained. Its further condensation reaction with 1,1-dimethoxy-N,N-dimethylethanamine (compound 11 in Fig.4) produces N,N-dimethylamino-vinyl group containing boron enolate (compound 12 in Fig.4). Then following an acidic treatment gives 2-tert-butyl-6-methyl-4H-pyran-4-one (compound 13 in Fig.4). However this method has a drawback because two synthetic reactions towards our target compound had low yields (30-40%), which results in a very low overall yield for synthesis of 2-tert-butyl-6-methyl-4H-pyran-4-one (compound 13 in Fig.4).
Fortunately, there is another method for synthesizing 2-tert-butyl-6-methyl-4H-pyran-4-one (compound 13 in Fig.4) with good yields  using pentane-2,4-dione (compound 14 in Fig.5) as starting reactant.
In its Aldol reaction with methyl pivalate (compound 15 in Fig.5) a 7,7-dimethyloctane-2,4,6-trione (compound 16 in Fig.5) was formed. Without separating the compound 16 from reaction mixture it was subjected to acidic cyclization producing 2-tert-butyl-6-methyl-4H-pyran-4-one (compound 13 in Fig.5) with a good overall yield (60%). As with 2-isopropyl-6-methyl-4H-pyran-4-one (compound 8 in Fig.3), the resulting 2-tert-butyl-6-methyl-4H-pyran-4-one (compound 13 in Fig.5) also contains one carbonyl group and one activated methyl group with the possibility of also adding only one aromatic aldehyde containing fragment.
One of oldest, but no less important methods known for the synthesis of 2,6-disubstituted-4H-pyran-4-ones is to obtain them from 3-substituted-vinylcarbonyl-4-hydroxy-6-methyl-2H-pyran-2-ones (compounds 17 in Fig.6) . Compounds 17 are obtained from dehydroacetic acid (compound 1 in Fig.6), in which the methyl group in the acetyl fragment is activated to react preferentially with aromatic aldehydes (see Fig.6) giving 3-substituted-vinylcarbonyl-4-hydroxy-6-methyl-2H-pyran-2-ones (compounds 17 in Fig.6) [33, 36]. Details on the obtained compounds 17 and their dependence on substituents (R) in their molecules are given in Table 1. They serve as precursors for further synthesis of pyranylidene type compounds.
Using this approach it is possible to obtain many different mono-styryl-substituted 4H-pyran-4-ones (compounds 17 in Fig.7). However only a few of previously synthesized compounds 17 give 2-styryl-substituted-6-methyl-4H-pyran-4-ones (compounds 18 in Fig.7) by acidic decarboxylation under the reaction conditions reported in [30, 33] (see Fig.7) as summarised in Table 2.
|R (of compounds 17)||Yield, %||M.p., °C||Recrystallized from|
|p-Dimethylaminophenyl||71||198-200||Chloroform, ethyl acetate, benzene|
|p-Diethylaminophenyl||58||150||Chloroform, ethyl acetate|
|3,4-Dichlorophenyl||46||185||ethyl acetate, benzene/chloroform|
|R (of compounds 18)||Yield, %||M.p., °C||Recrystallized from|
|p-Dimethylaminophenyl||82||156||ethyl acetale/petroleum ether|
Some works can be found on red luminescent compounds where the pyranylidene fragment is hidden as a substructure in the molecule [8, 23-24]. For example, chromene type derivatives of pyranylidene are synthesized from 1-(2-hydroxyphenyl)ethanone (compound 19 in Fig.8) [23-24]. In the Claisen condensation reaction (see Fig.8) with ethyl-acetate in the presence of a strong base, 1-(2-hydroxyphenyl)butane-1,3-dione (compound 20 in Fig.8) is obtained. After separation it was subjected to acidic dehydrocyclization giving 2-methyl-4H-chromen-4-one (compound 21 in Fig.8) with an overall 45% yield.
In the first stage of synthesis, treatment with morpholine gives us enamine 23 (4-(6-methylcyclohex-1-enyl)morpholine). In the second stage of synthesis in reaction with 2,2,6-trimethyl-4H-1,3-dioxin-4-one, a 2,8-dimethyl-5,6,7,8-tetrahydro-4H-chromen-4-one (compound 24 in Fig.9) is sucessfully obtained. Once the desired pyranylidene compound is obtained, the addition of electron acceptor and electron donor fragments becomes a more simplified process, which will be described in detail below in this chapter.
2.2. Addition of electron acceptor fragments to derivatives of 4H-pyran-4-ones and 3,5,5-trimethylcyclohex-2-enone
The next step towards synthesizing fully functional pyranilydene and isophorene type red luminescent organic compounds is the addition of electron acceptor fragments to the previously obtained 2,6-disubstituted-4H-pyran-4-ones (see Fig.10) and 3,5,5-trimethylcyclohex-2-enone (see Fig.11).
Many different electron acceptor fragments (compounds 25-35 in Fig.10) can be introduced in 2,6-disubstituted-4H-pyran-4-ones [1,4-18, 28-30,32] using acetic anhydride (Ac2O) as solvent and catalyst. From these, malononitrile (compounds 25 in Fig.10) is the most commonly used. Since isophorene (3,5,5-trimethylcyclohex-2-enone) (compound 36 in Fig.11) is an inexpensive reagent, which can be purchased from chemical suppliers - such as ACROS and ALDRICH, all that remains is to add electron acceptor and electron donor fragments. As with 2,6-disubstituted-4H-pyran-4-ones, the electron acceptor fragments are added in Knoevenagel condensation reactions [18-21, 31, 37] with active methylene group containing compounds 37-39 (see Fig.11).
The electron acceptor fragment containing derivatives of isophorene (3,5,5-trimethylcyclohex-2-enone) (compounds 37-39 in Fig.11) thus obtained are not always isolated from the reaction mixture [31, 37]. Once they are formed, the electron donor fragment containing aromatic aldehyde is added in the mixture for further reaction with the aldehyde.
2.3. Synthesis of pyranilydene and isophorene type red luminescent compounds by final addition of electron donor fragments
Once the electron acceptor fragment is introduced, the last step for obtaining a fully functional pyranylidene and isophorene red luminescent compounds is to add one or two electron donor fragment containing aldehydes. They are added in Knoevenagel condensation reactions with electron acceptor fragment containing derivatives of isophorene as shown in Fig.12 and pyranylidene shown in Fig.13, which contain one or two activated methyl groups.
For isophorene type compounds one electron donor fragment (40-44) is always introduced after an electron acceptor fragment is already in the molecule (see Fig.12) [18-21, 31, 37]. Many different structures of electron donor fragments are introduced (compounds 45-57 in Fig.13) in the pyranylidene backbone after introducing the electron acceptor fragment [1,4-18,27-29,31]. In cases where only one methyl group reacts with the aldehyde, a mono-styryl derivative of pyranylidene is obtained (see Fig.13). However, as all possible combinations shown in Fig.13 have not yet been synthesized, it presents a working opportunity for many organic chemists to contribute. If a pyranylidene type compound has two active methyl groups, like compound 25a, (see Fig.10) it will react with one or two aromatic aldehyde molecules producing chromophores 58-66 (see Fig.14). The reaction product will most likely be a mixture of mono- and bis- condensation products, which are difficult to separate and purify . In reaction with two methyl groups bis-styryl derivatives of pyranylidene are obtained (see Fig.14).
A good summary on dicyanomethylene-pyranylidene type red-emitters has been made by Guo et al. , according to which the mono electron donor fragment containing pyranylidene-type materials (45a-c to 57a-c in Fig.13) usually have high luminescence quantum yield but their chromaticity is not sufficiently good. At the same time two electron donor fragment derivatives of pyranylidene (compounds 58-66 in Fig.14) have better chromaticity, but their luminance efficiency is relatively low, particulary those with larger conjugations leading to a broad light-emission peak above 650 nm extending to the NIR region, which decreases the efficiency of red electroluminescent materials.
Both chromene (compounds 47,49-50 in Fig.13) and benzopyran (compounds 47,49,51 in Fig.13) type derivatives of pyranylidene have only one electron donor fragment in their molecules, but their optical properties are different. Since chromene type derivatives of pyranylidene have an additional conjugated aromatic ring in its molecule, its optical properties are similar to those with two electron donor fragment derivatives of pyranylidene (compounds 58-66 in Fig.14). At the same time benzopyran pyranylidene compounds 45,46,49 have a simple cyclohexene ring without additional conjugation, so their optical properties are more similar to pyranylidene-type red-emitters, compounds 45a-c to 57a-c.
If a pyranylidene backbone with different electron acceptor fragments contains two active methyl groups, then in reaction with a two aldehyde group containing compounds a polymer is formed during the reaction (see Fig.15) . The resulting polymers 70-72 are also reported to be red light-emitting materials.
All derivatives of pyranylidene and isophorene reported so far in this chapter are deposited on the OLED hole transport layer either by thermal evaporation in vacuum or used as dopants in a polymer matrix in limited concentrations.
3. Synthesis and properties of trityloxy group containing glassy derivatives of pyranylidene and isophorene
Our key for obtaining glass forming materials is the synthesis of such electron donor substituent containing aldehyde which would ensure the formation of an amorphous structure of our newly synthesized derivatives of pyranylidene and isophorene. We have synthesized such a compound - 4-(bis(2-(trityloxy)ethyl)amino) benzaldehyde [31-32] 75, in Fig.16.
3.1. Preparation of molecular glasses
For obtaining a red luminescent glass forming derivative of isophorene, we start with (3,5,5-trimethylcyclohex-2-enone) (compound 29 in Fig.16) as already described in Fig.9. It is subjected to the Knoevenagel condensation reaction with malononitrile (28). However, 2-(3,5,5-trimethylcyclohex-2-enylidene)malononitrile (61) which is formed during the reaction is not isolated because 4-(bis(2-(trityloxy)ethyl)amino) benzaldehyde (75) is added to the reaction mixture after 2 hours [31, 37] for further reaction. 2-(3-(4-(Bis(2-(trityloxy)ethyl)amino)styryl)-5,5-dimethylcyclohex-2-enylidene)malononitrile (IWK) was obtained in good yield after its separation and purification by liyquid column chromatography as described in .
For obtaining red luminescent glass forming derivatives of pyranylidene, we use three different electron acceptor fragment containing derivatives of pyranylidene (compounds 25a in Fig.17). Malononitrile (in compounds 74a and 75a), indene-1,3-dione (in compounds 74b and 75b) and barbituric acid (in compounds 74c and 75c) are used as electron acceptor fragment carrying compounds .
In the Knoevenagel condensation reaction with compound 25a and 4-(bis(2-(trityloxy)ethyl)amino) benzaldehyde (73) a mixture of mono- (ZWK-1, DWK-1, JWK-1) and bis- (ZWK-2, DWK-2, JWK-2) condensation products is obtained. Their separation is complicated but nevertheless a large part of each product was separated by liquid column chromatography (silicagel and dichloromethane for ZWK-1 and ZWK-2, dichloromethane: hexane = 4:1 for DWK-1 and DWK-2, dichloromethane: ethyl acetate = 4:1 for JWK-1 and JWK-2). The physical properties of compounds WK-1, WK-2 and IWK are described in detail further in this chapter.
3.2. Thermal properties
The thermogravimetric analysis (TGA) of trityl group containing pyranylidene type compounds is used to measure their thermal decomposition temperatures (Td). Td of compounds WK-1 and WK-2 are determined in the temperature range from +30°C to +510°C at a heating rate of 10°C/min  at the level of 10% weight loss (see Fig.18).
Pyranylidene type compounds with two N,N-ditrityloxyethylamino electron donor fragments (ZWK-2, DWK-2, JWK-2) are slightly more thermally stable than compounds containing only one such fragment, i.e. ZWK-1, DWK-1 or JWK-1. The increase in thermal stability of pyranylidene type compounds by adding another electron donor fragment is as high as 10°C from ZWK-1 to ZWK-2, 19°C from JWK-1 to JWK-2 and 29°C from DWK-1 to DWK-2. The most thermally stable compound is a two electron donor fragment containing derivative of pyranylidene with malononitrile as electron acceptor in it (DWK-2).
Differential scanning calorimetry (DSC) measurements are used to measure the glass transition temperatures (Tg) of the compounds WK-1 and WK-2. Three thermo cycles are performed for the determination of Tg. The first scan was done within the temperature range
from +25°C to +250°C at a heating rate of 10°C/min . After the first heating scan samples of the compounds were cooled to 25°C at a rate of 50°C/min and heated for a second time from +25°C to +250°C at a rate of 10°C/min. The Tg value is obtained from the second heating scan (see Fig.19) and for almost all compounds is higher than 100°C. We could not obtain usable DSC curves for DWK-1. The compounds with two N,N-ditrityloxyethylamino electron donor fragments have higher Tg compared to those with only one electron donor fragment, which may be attributed to the different numbers of bulky trityloxyethyl groups attached to the two electron donor fragment. In a larger number of bulky groups Tg increases by 8°C from ZWK-1 to ZWK-2 and 7°C from JWK-1 to JWK-2. Pyranylidene type compounds with barbituric acid as electron acceptor, e.g. JWK-1 and JWK-2 have the highest Tg values compared to ZWK-1, ZWK-2 and DWK-2, which may be due to the additional formation of intermolecular hydrogen bonds by N-H groups of barbituric acid fragments in the molecules.
The TGA analysis of IWK is conducted as previously described . The thermal decomposition temperature (Td) of IWK is found to be even higher than that of pyranylidene type compounds WK-1 and WK-2 (see Fig.20). However its glass transition temperature (Tg) is lower by 18°C to 35°C degrees compared to that of pyranylidene type glasses. Despite the lower thermal stability, the pyranylidene type compounds WK-1 and WK-2 have better glass forming properties than the isophorene type compound IWK.
3.3. Glass forming properties
Thin films are deposited on quartz glass by the spin-coating technique. Before the deposition of the layers, the quartz glass substrates are cleaned in dichloromethane. The solutions are spin-coated onto the substrates for 40 s at 400 rpm and acceleration 200 rpm/s.
In all cases, pure films obtained from two electron donor fragment containing pyranylidene compounds (ZWK-2, DWK-2 and JWK-2) have an almost pure smooth and amorphous surface, but pyranylidene compounds with one electron donor fragment (ZWK-1, DWK-1 and JWK-1) show several crystalline state areas (see Fig.21). Both glasses containing barbituric acid as an electron acceptor fragment (JWK-1 and JWK-2) show the least amount of small crystal formations on their pure film surface. The higher stability of their amorphous state could be explained by an enchancement of N-H group hydrogen bonds in the molecules. Pure films obtained from malononitrile electron acceptor fragment containing compounds (DWK-1 and DWK-2) contain small crystal dots, especially DWK-1. This could be due to small steric dimensions of malononitrile group, which allows more DWK-1 molecules to be concentrated in the same volume to allow closer interaction with other molecules enabling higher possibility to form agreggates and crystallites.
Information obtained from the surfaces of the pure films is consistent with the measured glass transition temperatures (Tg). Glasses having higher Tg values are found to have less crystalline dots on their pure film surface. As we were unable to determinate Tg for DWK-1, according to above mentioned trend its glass transition temperature is expected to be below 110°C.
Thin film containing only pyranylidene type compound WK-1 and WK-2 are amourphous despite of small crystalline dots in it. Till now only way to prepare amourphse films which contain pyranyliden derivatives was doping them in glass forming compound. In that case maximum doping concentration was considered to be 2wt% due to self crystallization [11-12]. However, incorporation of bulky trityloxy groups in their molecules or using glasses WK-1 and WK-2 could increase this concentration limit more then 10 times.
3.4. Absorption and luminescence properties
A DWK-1 molecule, whose backbone consists of the laser dye 4-(dicyanomethylene)-2-methyl-6-[p-(dimethylamino)styryl]-4H-pyran (DCM), in dichloromethane solution has its absorption maximum at 472 nm, which is 8 nm red shifted with respect to the pure DCM molecule in the same solution . It shows that the bulky trityloxyethyl group has only a small influence on the energy structure of the molecule. The peaks of the absorption spectra in solution of the molecules with indene-1,3-dione (ZWK-1) and barbituric acid (JWK-1) electron acceptor substituents in the backbone are red-shifted by approximately 40 nm compared to DWK-1. A stronger electron acceptor group gives larger red shifts.
The photoluminescence (PL) spectrum of the DWK-1 solution was found to be Stokes shifted by about 115 nm (peak position at 587 nm) with respect to the absorption spectra (see Fig.22). The PL spectra of JWK-1 and ZWK-1 molecules exhibited similar shapes, with their maxima red-shifted to 635 and 627 nm, respectively. The photoluminescence spectra are unstructured and strongly Stokes shifted in accordance with intramolecular charge-transfer nature of the excited states . For compounds containing two 4-((N,N-ditrityloxyethyl) amino)styryl electron donor fragments the absorption and luminescence spectra of the solution are observed to be red-shifted and have larger extinction coefficients, which is due to the larger absorption cross section of these molecules. The peaks of the absorption spectra of DWK-2 and ZWK-2 are red shifted by 17 and 11 nm, respectively, compared to molecules with a single electron donor fragment. A similar red shift has been reported for the molecule with two electron donor fragments bis-DCM compared with DCM molecules with a single electron donor fragment . It is observed that molecules with two electron donor fragments have a larger conjugation length. A second reason could be simultaneously functioning two donor groups which give stronger electron donor properties. The shape of the absorption spectrum of JWK-2 is found to be different from that of JWK-1 and the oscillator strength of the absorption band of JWK-2 at about 502 nm becomes more intense (see Fig.22(1)).
The fluorescence spectra of molecules with two electron donor fragments are broader and further Stokes shifted than molecules with only one electron donor fragment. This may be attributed to the different conjugation lengths as indicated by the absorption spectra. The peak positions of DWK-2, ZWK-2 and JWK-2 are observed at 640, 678 and 701 nm, respectively. The red shift of the absorption spectra of solutions increases corresponding sequentially to ZWK, JWK and DWK, as stronger electron acceptor fragments induce larger red shifts. This could be explained by their electron withdrawing properties, which differ among our investigated electron acceptor fragments. The shift of luminescence spectra did not maintain the same sequence due to the larger Stokes shift for the JWK molecules.
The absorption spectra of thin solid films of the molecules with one electron donor fragments are practically unchanged with respect to the solutions spectra. They are slightly broader with small red-shift indicating a weak excitonic interaction in the solid state, which is typical for glass-forming amorphous materials. For the molecules with two electron donor fragments ZWK and DWK the absorption spectra are found to shift by 21 nm and 22 nm, respectively. The peak positions of the absorption spectra for JWK molecules remain unchanged by the incorporation of a second electron donor fragment. However, the fluorescence spectra of all films are red-shifted in comparison with those of solution.
For molecules with one electron donor fragment, the shape of the fluorescence spectra of thin films is very similar to that in solution, which confirms that for these compounds the excited states in the aggregates in the solid state are not very different from those in molecules. However, the derivatives with two electron donor fragments exhibit an additional band at longer wavelengths in thin films, which becomes more intense going from weaker to stronger electron acceptor fragments in the studied molecules. In the case of ZWK-2 in thin films the additional band even becomes dominant.
In the case of IWK, the absorption and luminescence spectra of thin solid films are also found to be practically unchanged compared to its solutions spectra as shown in Fig.24.
The same relation is observed for IWK emission properties in solution as well as in thin films. However, in solid state its emission is very weak compared to pyranylidene type compounds, which may limit the usefulness of IWK in OLED applications.
3.5. Photoluminescence quantum yields
Photoluminescence quantum yield (PLQY) of the investigated compounds in solution and in thin films is measured by using an integrating sphere (Sphere Optics) coupled to a CCD spectrometer . PLQY thus measured for all compounds are summarised in Table 3. Compounds with more polar groups attached exhibit PLQY up to 0.54 in dilute solutions, which is slightly higher than for DCM dye in similar surroundings [42, 43]. PLQY depends slightly on the acceptor group as can be seen from Table 3. That means that compounds with a stronger electron acceptor group have higher PLQY. JWK and ZWK molecules with two electron donor groups have lower PLQY in comparison with one electron donor group. However, the opposite is observed with DWK compounds, as molecules with two electron donor groups exhibit larger PLQY. This may be due to the shielding of the acceptor group by bulky trityloxyethyl groups. PLQY of pure films is found to be more than one order of magnitude lower than that in solution. This reduction is particularly strong in the case of molecules with two donor groups. PLQY values of these compounds correlate with the intensity of the long wavelength fluorescence band, as PLQY is lower in materials with a stronger low energy fluorescence band. Molecular distortions taking place during formation of solid films are probably responsible for both of these effects. Compound molecules with two bulky acceptor groups are probably strongly distorted in solid films, so that molecular chains connecting acceptor and donor moieties are twisted. Such twisting usually leads to a red-shift in the molecular fluorescence and to fast non-radiative relaxation . The twisted molecules form energy traps in solid films, which may be populated during the excitation diffusion. Therefore, even a small fraction of distorted molecules may significantly affect the fluorescence spectrum and PLQY. We were unable to measure PLQY in IWK pure thin solid films. Moreover, it also shows the lowest value in solution and therefore cannot be used as a light-emitting material.
It is worth mentioning that DCM molecules do not show any photoluminescence from pure films due to the small distance between molecules which results in high molecular interaction. Therefore, host-guest films of transparent polymethylmethacrylat (PMMA) polymer with varying dye doping were prepared in order to observe the impact of concentration on photoluminescence quenching. The dependence of PLQY on concentration of DWK-1 and DWK-2 molecules is shown in Fig.25.
For comparison the PLQY of DCM in PMMA are also included in Fig.25. PMMA films doped with DWK-1 and DWK-2 at low concentration (<1 wt%) exhibit somewhat lower PLQY as compared to that obtained in solution (See Fig.24 and Table 3). This discrepancy may be attributed to the sensitivity of molecules to the polarity of the surrounding media. At higher concentrations (>3 wt%) the DWK-1 molecule shows negligible photoluminescence quenching dependence on concentration. On the other hand molecules with two donor groups exhibit pronounced quenching. Fluorescence efficiency of the polymer film doped with 10 wt% of DWK-2 molecules decreases 2-times compared to that of films doped with 10 wt% DWK-1 molecules. The reason for the lower PLQY could be the same as for different PLQY of the pure films. The laser dye DCM dispersed in the polymer matrix at high concentration shows a remarkable fluorescence quenching. For example, at a 10 wt% concentration of DCM molecules, up to a 4-time decrease of quantum yield is observed in comparison with the same concentration of DWK-1 molecules. Thus, incorporation of bulky trityloxyethyl groups prevents the formation of aggregates of the dye molecules and remarkably reduces the fluorescence quenching dependence on concentration, enabling the use of higher doping levels in emissive layers.
3.6. Amplified spontaneous emission properties
In order to test the light amplification prospects of our synthesized compounds, we prepared pure thin films of all the compounds on a quartz substrate and measured their amplified spontaneous emission (ASE). Such emission was observed only for four of six compounds, DWK-1, DWK-2, JWK-1 and ZWK-1, as shown in Fig.27 .
From the other two samples of JWK-2 and ZWK-2 no ASE signal has been observed. The peak positions of ASE are red shifted as compared to the fluorescence band maxima (see Fig.27 and Fig.22). The red shift values were found to be 14, 18, 10 and 31 nm for DWK-1, DWK-2, JWK-1 and ZWK-1, respectively. Variations in the peak intensity of ASE spectra as a function of the pump beam pulse energy are shown in Fig.28, from which ASE threshold values are estimated to be 90±10, 330±20, 95±10, 225±20 μJ/cm2 for DWK-1, DWK-2, JWK-1 and ZWK-1, respectively. These values are larger in comparison with the threshold values (of the order of micro joules per square centimeter) reported for some other materials [46, 47].
However, a direct comparison is difficult because the ASE threshold, in addition to material properties, depends also on the sample and excitation geometries, film thickness, optical quality and excitation pulse duration.
Nevertheless it has not been observe ASE in pure DCM films, but we have measured it in DWK-1 which is the same DCM with additional trityloxyethyl group. It should also be noted that some sample degradation has been observed at the highest excitation intensities; however no noticeable degradation is observed when excitation intensity is 1.5 - 2 times exceeding the ASE threshold.
ASE develops in the spectral position where the light amplification coefficient has the maximal value. The amplification coefficient may be described as:
where n* is the density of excited molecules, N is the total density of molecules, σ0(λ),σem(λ)and σ*(λ) are cross-sections of the ground state absorption, stimulated emission and excited state absorption, respectively. As it can be seen from Eq. (1) even weak ground state absorption may strongly reduce the amplification coefficient or make it negative. This is because only a small fraction of molecules is usually excited even under high intensity excitation conditions, i.e., N>>n*. Thus, the absorption band tails, which overlap with fluorescence band, are evidently responsible for the red shifts in ASE spectra in comparison with the maxima of the fluorescence. Note, that the light propagation length is limited by the film thickness in the absorption measurements, while ASE emission can propagate a much longer way along the film.
3.7. Photoelectrical properties and energy structure of glassy thin films
Information about the location of energy levels enables one to determine the best sample structure for electroluminescence measurements. To characterize the energy gap in organic solids several methods are applied. In organic crystals as well as amorphous solids charge carriers do not emerge as “bare” quasi-free electrons and holes but as a polaron type quasi-particle, dressed “in electronic and vibronic polarization clouds” [48, 49]. Electronically relaxed charges may be formed far enough from each other which give rise to a wider optical band gap EGOpt [49, 50]. The optical energy gap EGOpt may be obtained from the low energy threshold of the absorption spectra of organic thin films. The vibrationally and electronically relaxed charge carrier states contribute to the adiabatic energy gap EGAd. It could be attributed to the threshold energy of photoconductivity Eth which can be estimated from the spectrum of the quantum efficiency of photoconductivity β(hυ) :
where jph is the density of photocurrent at a given photon energy hν and applied voltage U, I(hν) is the intensity of light (photons/cm2s), k(hν) is the transmittance of the semitransparent electrode and g(hν) is the coefficient which characterizes the absorbed light in the organic layer.
Eth can be determined from a sample where the organic compound is sandwiched between two semitransparent electrodes, which in our case are ITO and thermally evaporated aluminum. The sample is irradiated through the electrodes and current changes are measured as shown in Fig.29(1). Efficiency of photoconductivity at different light energy is calculated using Eq. (2) and is plotted as a function of the photon energy in Fig.29(2). The sample is illuminated from both aluminum and ITO side when positive and negative voltage is applied to them. Eth is determined by plotting β2/5 as a function of the photon energy. The intersections of tangents at low photon energy on the curve of β2/5 plotted as a function of the photon energy and photon energy axis gives Eth as shown in Fig.29(3).
Optical band gap EGOpt, photoconductivity threshold value Eth and reduction-oxidation potential Uredox, determined from cyclic voltamperogramme, for investigated compounds are presented in Table 4.
|EGOpt (eV)||Eth (eV)||Uredox (V)|
According to Table 4, the redox potential of DWK, JWK and ZWK is higher for compounds with one electron donor group compared to compounds with two electron donor groups (see Fig.17). The same relation is found for optical band gap as well.
The photoconductivity threshold value cannot be obtained for DWK-2 thin films due to the low value of photocurrent. For other compounds we obtain an excellent linear correlation between optical band gaps and photoconductivity threshold values with correlation coefficient 0.993. The slope of this linear relation is found to be 1 and intercept 0.28 as shown in Fig.31
The energy of the photoconductivity threshold is defined as the difference between the conduction levels of holes and electrons . The value of the intercept implies that the optical band gap is 0.28 eV larger than the difference between the conduction levels of holes and electrons. It shows a constant energy difference between optical band gap and adiabatic gap despite the various molecule structures.
3.8. Electrical properties
Electrical properties of WK-1 and WK-2 compounds are investigated in the regime of space charge limited current (SCLC) [52-54]. Similar sandwich type samples as used for photoelectrical measurements are prepared for this study as well. The thickness of the organic thin film is at least 500 nm.
The current-voltage characteristics of compounds ZWK-1, ZWK-2, JWK-1, JWK-2 and DWK-2 in thin solid films are shown in Fig.32.
The current-voltage characteristic of DWK-1 films could not be measured due to unstable current. This may be due to formation of small crystallites (see Fig.21) around 1 μm in size. Such aggregates are found throughout the sample and induce instability in the current. In all other cases the current-voltage characteristics have similar shapes with three regions. In the first region, 0-2 volts, the current is found to depend linearly on voltage. In the second range, 2 to 50 volts the current increases superlinearly with voltage, following Child’s law. In the third region, > 50 V, the current depends on voltage to the power of at least ten, which may be attributable to charge trapping in the local trap states. More details of this aspect will be discussed further below.
Usually the work function of ITO should be near the ionisation energy level of the organic compound while that of aluminium (Al) should be around the middle of the energy gap. This provides efficient hole injection from ITO and electron injection from aluminium when a positive voltage is applied to ITO. Holes may also be injected from the aluminium when positive voltage applied to it. Electron injection may be more difficult in the second case due to the large difference between the ITO work function and electron affinity potential of the organic compound. This is confirmed by the current voltage characteristics shown in Fig.32. A similar current is observed at the lower voltage where only holes are injected either from ITO or aluminium when biased with a positive voltage. At higher voltage current is higher when ITO is positive in comparison with positive aluminium.
The temperature modulated space charge limit current (TM SCLC) method is used to analyse the charge carrier local trapping states in solid films . The condition for using this method (TM SCLC) is monopolar injection, which is achieved in our case when a positive voltage is applied to the aluminium electrode. The measured activation energy is plotted as a function of the applied voltage for the investigated compounds as shown in Fig.33.
No charge carrier local trap states are found in films of compounds with one electron donor group due to only one plateau which reaches zero. All compounds with two electron donor groups are found to have charge carrier trap states. The additional plateau of activation energy, which can be clearly seen from Fig.33 means that the thin films contain local trap states. The hole shallow trap depths are found to be 0.1, 0.24 and 0.3 eV in ZWK-2 JWK-2 and DWK-2, respectively. Such trap states decrease the efficiency of electroluminescence and should be avoided in fabricating high efficiency light emitting diodes. The activation energy increases at lower voltage for compounds JWK-1, JWK-2 and DWK-2. This is indicative of a contact problem where the electrode – organic interface also works as additional charge carrier traps.
3.9. Electroluminescence of ZWK-1 and ZWK-2
A multilayer structure is used for electroluminescence (EL) measurements. Polyethylenedioxythiophenne:polystyrenesulfonate (PEDOT:PSS) (from H.C. Starck) is used as the hole injection layer and LiF as electron injection layer. PEDOT:PSS and organic compounds are sequentially spin coated on ITO glass. Then LiF and Al are thermally evaporated in vacuum. The final structure of the device has a structure of ITO/PEDOT:PSS(40nm)/ZWK1 or ZWK-2(~90nm)/LiF(1nm)/Al(100nm) and is not encapsulated.
The EL spectrum of the device is estimated in International Commission on Illumination (CIE) coordinates: x=0.65 and y=0.34 for ZWK-1 and x=0.64 and y=0.36 for ZWK-2. The spectral maximum peak is observed at 667 nm and 705 nm in ZWK-1 and ZWK-2, respectively, as shown in Fig.34. These peaks are slightly red shifted compared with those of PL spectrum of ZWK-1 and ZWK-2 thin films. This red shift may be attributed to the interaction of molecules and injected charges.
The light emission is observed at 6 V in the electroluminescent device with ZWK-1 molecules and 9 V in with ZWK-2 molecules. The light intensity is one order less in ZWK-2 molecules compared to that in ZWK-1. This may be due to the lower PLQY and shallow charge carrier trap states in ZWK-2.
The absorption and emission bands of the synthesized pyranylidene type compounds ZWK-1, DWK-1, JWK-1 are comparable with those of other already known one electron donor fragment DCM and benzopyran type derivatives of pyranylidene within the spectral region studied here. Similar conclusions can be drawn about ZWK-2, DWK-2, JWK-2, which have similar properties to IWK and two other already known electron donnor group containing derivatives of pyranylidene. These properties are also similar to those of one electron donor fragment chromene red-emitters. However, incorporation of bulky trityloxy groups in such molecules not only enchances glass transition temperatures by 5° to 20°C compared to previously published pyranylidene type compounds containing one and two electron donor groups, but also enables the formation of a glassy structure in the solid state from volatile organic solvents. In addition, no glass transition values have been observed so far for low molecular mass isophorene type compounds. The photoluminescence quantum yield of investigated molecules in solution is up to 0.54 and is also comparable with the quantum yield of pyranylidene and isophorene derivatives already reported. Most of the thin solid films obtained from WK-1, WK-2 have almost no crystals in the sample. Newertheless the photoluminescence quantum yield is reduced by one order of magnitude due to the closer intermolecular distance between molecules, resulting in strong excitonic interaction. Emission from the IWK film is too weak to detect, which may be attributed to the higher photoluminescence quenching in IWK than in glassy pyranylidene films. However, using the doping approach, the compounds we have introduced enable up to 3 times higher doping concentration without losing optical properties compared to other already known red-emitters.
Four investigated compounds - ZWK-1, JWK-2, DWK-1 and DWK-2 show amplified spontaneous emission from pure solid films. Obtained threshold values are larger than those previously reported, but it should be mentioned that for pyranylidene type compounds, amplified spontaneous emission has been observed only in the doped systems until now.
Electrical properties are found to be better in compounds with one electron donor group due to absence of local trap states in their thin films. In the case of molecules with two electron donor groups shallow hole trap states have been observed, which may decrease efficiency of electroluminescence and should therefore be avoided in fabricating high efficiency light emitting diodes.
Even though we are able to prevent pyranylidene and isophorene type red-emitters from self crystallization in the solid state, their concentration in the emission layer would still be limited due to photoluminescence quenching caused by the short distance between molecules. Nevertheless, the glass materials can still be used not only as dopants for OLED applications, but also for lasing applications. Good thermal properties present a possibility of using them also for nonlinear optical (NLO) property studies.
This work has been supported by the European Social Fund within the project «Support for the implementation of doctoral studies at Riga Technical University» and «Support for Doctoral Studies at University of Latvia» and by Latvian State Research Programm No.2 in Materials Sciences and Information Technologies. Authors are grateful to Janis Jubles for assistance on reactant synthesis, Kristine Lazdovica for absorption and emission measurements in solutions and conducting thermogravimetric analysis, Dr. chem. Baiba Turofska for electrochemical measurements, Kaspars Pudzs for activation energy measurements, Raitis Grzibovskis for photoelectrical measurements, Dr. phys. Saulius Jursenas for photoluminescence quantum yield measurements, Dr. phys. Vidmantas Gulbinas for useful discussion and Dr. phys. Mikelis Svilans for providing language help.