Luminescence Quantum Efficiency of Nanocrystalline ZnS: Mn2+. 2

In this paper the influence of UV irradiation on the luminescence quantum ... of the luminescence quantum efficiency upon irradiation, due to UV curin...
0 downloads 0 Views 117KB Size
J. Phys. Chem. B 2001, 105, 10203-10209

10203

Luminescence Quantum Efficiency of Nanocrystalline ZnS:Mn2+. 2. Enhancement by UV Irradiation Ageeth A. Bol* and Andries Meijerink Debye Institute, Physics and Chemistry of Condensed Matter, Utrecht UniVersity, P.O. Box 80 000, 3508 TA Utrecht, The Netherlands ReceiVed: February 27, 2001; In Final Form: June 26, 2001

In this paper the influence of UV irradiation on the luminescence quantum efficiency of nanocrystalline ZnS:Mn2+ is studied. Samples passivated with poly(vinylbutyral) (PVB), poly(vinyl alcohol) (PVA), methacrylic acid (MA), and sodium polyphosphate (PP) and samples without passivating polymer were UV irradiated in air, nitrogen, dry air, and wet nitrogen. Samples coated with PVB, MA, and PVA show the highest increase of the luminescence quantum efficiency upon irradiation, due to UV curing of the passivating polymer. The UV enhancement observed for the PP-coated sample and the unpassivated sample is explained by photochemical reactions that take place at the surface of the nanoparticles during the irradiation. The products of these reactions (e.g., ZnSO4 or Zn(OH)2) serve as a passivating layer around the nanoparticles, which increases the quantum efficiency. The highest steady-state quantum efficiencies obtained after UV curing are around 10%. In addition to the long-term UV enhancement of the Mn2+ emission intensity, short-term (seconds) UVinduced quenching of the Mn2+ emission is reported and a model for the quenching is discussed.

1. Introduction During the past two decades the interest in nanocrystalline semiconductors has increased rapidly. The changes in the electronic structure as a function of particle size (quantum size effects) are intriguing and have stimulated fundamental research on various types of nanocrystalline semiconductors.1-5 Not long ago, reports on nanocrystalline semiconductors doped with luminescing ions have appeared.6-9 Efficiently luminescing nanocrystalline ZnS:Mn2+ particles may be interesting for application in, e.g., electroluminescent devices. Hence, it is important to study which factors influence the luminescence quantum efficiency of these nanoparticles. Bhargava6 reported that with decreasing particle size the quantum efficiency increases from 1% for particles of 7 nm to 18% for particles of 3.5 nm due to quantum size effects. Other researchers10-13 have found that the quantum efficiency of nanocrystalline ZnS:Mn2+ is dependent on the Mn2+ concentration. Also the influence of capping polymers on the luminescence quantum efficiency has been investigated,13-15 and in addition a significant increase of the luminescence quantum efficiency due to UV irradiation has been observed.15-21 In addition to relative changes in the quantum efficiency, it is important to measure absolute quantum efficiencies. Sometimes a very large relative increase in quantum efficiency is reported (e.g., 1000%21), but when the initial and final quantum efficiencies are not mentioned, this number is not of great significance. It will be much easier to increase a quantum efficiency of very poorly luminescent samples from 0.1% to 1% than to increase a quantum efficiency from 1% to 10%. To obtain a better understanding of factors which influence the quantum efficiency of nanocrystalline ZnS:Mn2+, we have performed systematic research on the influence of Mn2+ concentration, passivating polymer, and UV irradiation on the * Corresponding author. Fax: +31 30 253 2403. E-mail: a.a.bol@ phys.uu.nl.

absolute quantum efficiency. Experiments were repeated to test the reproducibility. In this paper the effect of UV irradiation is analyzed. In a preceding paper13 the roles of the Mn2+ concentration and the type of passivating polymer are reported. To explain the increase in quantum efficiency upon UV irradiation, several mechanisms have been proposed. Becker and Bard16 explained this phenomenon by oxygen absorption (perhaps as an oxide) upon irradiation, which effectively blocks (nonradiative) surface state recombination. This increases the luminescence quantum efficiency. In the mechanism proposed by Henglein17 oxygen also plays a role. During irradiation in the presence of oxygen, photoanodic dissolution of the ZnS nanoparticles takes place. The decrease in particle size was supposed to cause an increase of the luminescence quantum efficiency of the nanoparticles. Dunstan et al.18 explained the enhancement of the luminescence of nanocrystalline ZnS upon illumination by a photocorrosion process, in which products are formed, which could serve as new recombination centers. Later, Yu’s group19,20 proposed a model in which the efficiency of energy transfer from the ZnS host to the Mn2+ impurity increased with UV exposure. This was called the irradiation induced light enhancement effect (IILE effect). A similar mechanism was later proposed by Isobe et al.15 Efficient energy transfer from the polymer to the nanoparticles is suggested as being responsible for the increase in quantum efficiency. Finally, Gallagher and Bhargava21 explained the UV enhancement by further cross-linking and polymerization of the passivating molecules (UV curing) at the surface of the nanoparticles, which results in a better passivation of surface states that act as nonradiative recombination centers. All these proposed mechanisms for the UV enhancement evoke a few questions. For example, one wonders if one mechanism is true or if there is a combination of mechanisms. In addition, no absolute quantum efficiencies were measured in most of the above-mentioned papers. This makes it very hard to compare the reported results. To gain more insight in the

10.1021/jp010757s CCC: $20.00 © 2001 American Chemical Society Published on Web 09/28/2001

10204 J. Phys. Chem. B, Vol. 105, No. 42, 2001

Bol and Meijerink

mechanisms that cause UV enhancement, we investigated the time evolution of the luminescence of nanocrystalline ZnS:Mn2+ upon UV illumination in a systematic way. Several polymers were used to passivate the nanoparticles, and the atmosphere in which the samples were illuminated was varied. Absolute quantum efficiencies were measured as well. We have found that the UV enhancement of the various samples can be explained by two mechanisms: UV curing of the passivating polymer and passivation by the photooxidation of the surface of the nanoparticles. The atmosphere in which the samples are UV irradiated plays an important role. 2. Experimental Section The synthesis route followed to make nanocrystalline ZnS: Mn2+ resembles standard methods for synthesis of nanocrystalline II-VI semiconductors. The method used for the synthesis of nanocrystalline ZnS:Mn2+ coated with PP is very similar to the one described by Yu et al.,22 except that the synthesis was done in water instead of methanol and is described in detail in a preceding article.13 X-ray powder diffraction patterns of the nanoparticles were measured with a Phillips PW 1729 X-ray generator using Cu KR radiation (λ ) 1.542 Å). From the line width the particle diameter was calculated using the Scherrer formula.24 Emission and excitation spectra were recorded on a SPEX Fluorolog spectrophotometer Model F2002, equipped with two double grating 0.22 m monochromators (SPEX 1680) and a 450 W xenon lamp as a excitation source. The emission was detected with a cooled Hamamatsu R928 photomultiplier. Emission spectra were corrected for the sensitivity of the photomultiplier tube and monochromator characteristics, and excitation spectra were corrected for the intensity of the xenon lamp transmitted by the excitation monochromator. The quantum efficiencies of the various samples were calculated using the phosphor BaMgAl10O17:Eu2+ (BAM, commercial phosphor purchased from Philips) as a standard. The procedure is described in detail in ref 13. To ensure the same packing density, both the phosphor and the sample were weighed beforehand. The UV enhancement of the luminescence intensity was studied in several ways. During the UV irradiation, which was carried out in the spectrofluorometer at an excitation wavelength of 300 nm, emission and excitation spectra were measured as a function of irradiation time. The UV enhancement experiments were carried out in different atmospheres, namely, air, nitrogen, dry air, wet nitrogen, and ethanol. Before every measurement the sample chamber was flushed with nitrogen for 1/2 h. Then the sample chamber was flushed for 10 min with the gas in which the irradiation was going to take place. Dry air was created by bubbling air through concentrated sulfuric acid. In this way the water from the air was removed. Wet nitrogen was produced by bubbling nitrogen through water. An ethanol atmosphere was created by bubbling nitrogen through ethanol. 3. Results and Discussion 3.1. UV Irradiation of Nanocrystalline ZnS:Mn2+ Passivated with Different Polymers. Previous studies have reported a significant increase in quantum efficiency for UV-irradiated ZnS:Mn2+ nanoparticles. Increases of the luminescence of nanocrystalline ZnS:Mn2+ with a factor of 2.5 (samples coated with acrylic acid),15 3.5 (samples coated with PVB),19,20 10 (samples coated with PMMA (poly(methyl methacrylate))),21 and even 30 (samples coated with 3-methacryloxypropyltrimethoxysilane)24 were reported in the literature. The differences in increase percentage could possibly be explained by the fact

Figure 1. Emission spectra (λexc ) 300 nm) of nanocrystalline ZnS: Mn2+ passivated with MA measured after different irradiation times in air at 298 K.

that in every paper another polymer was used to passivate the nanoparticles. Since no absolute quantum efficiencies are measured in any of these articles, it is hard to compare the various results. To get more insight into the influence of the polymer on the UV enhancement, samples coated with different polymers were irradiated. Absolute quantum efficiencies were measured as well. Figure 1 shows several emission spectra (λexc ) 300 nm) after different irradiation times for a sample coated with MA. For all the other samples similar changes in the emission spectra were found. The broad purple/blue emission around 420 nm, which is also observed in undoped nanocrystalline ZnS, can be assigned to a defect-related emission of the ZnS. The orange emission around 590 nm can be attributed to the 4T1-6A1 transition of the Mn2+ ion. Both emission bands increase in intensity during UV irradiation. The relative increase of the ZnSrelated emission is stronger than the relative increase of the Mn2+ emission. Probably, the defect emission mainly originates from ZnS particles in which no Mn2+ ions are incorporated. The defect emission has been assigned to radiative recombination of a delocalized charge carrier with a trapped charge carrier.25 For the delocalized charge carrier, trapping in surface states is efficient and competes with radiative recombination. Passivation of surface states will strongly increase the relative contribution of radiative recombination. For the Mn2+emission, trapping of the charge carriers by Mn2+ (which is very fast) competes with trapping of charge carriers in surface states (followed by nonradiative relaxation). Once the excitation energy is localized on a Mn2+ ion in the excited state, the surface passivation is less important. This model can qualitatively explain the stronger relative increase of the defect-related emission. The small shift of the maximum of the orange emission band upon UV irradiation is probably due to the increase of the intensity of the UV/blue emission band. When the intensity of the ZnS-related emission or the Mn2+ emission is followed in time during the UV irradiation, curves such as the one shown in Figure 2 are obtained for all samples. After a substantial rise in the beginning the intensity levels off and reaches a constant value. It was observed that the UV enhancement behavior (e.g., initial rise, time to level off, total increase) is dependent on the polymer used to passivate the sample. Table 1 shows typical results of the increase percentages of the luminescence intensity (Increase%) and the quantum efficiencies before irradiation obtained for samples coated with

Luminescence Quantum Efficiency of ZnS:Mn2+. 2

J. Phys. Chem. B, Vol. 105, No. 42, 2001 10205

Figure 2. Luminescence intensity at λem ) 590 nm (λexc ) 300 nm) as a function of UV-irradiation time of nanocrystalline ZnS:Mn2+ passivated with MA measured in air.

TABLE 1: Increase Percentage (Increase%) of the Luminescence and Quantum Efficiencies before Irradiation (QEbefore) of Samples Coated with Various Polymers polymer

QEbefore (%)

Increase%

PVB MA PVA PP without

3.5 1.0 2.7 4.5 0.3

250 240 230 60 100

different polymers. The Increase% is defined as the quantum efficiency after UV irradiation (QEafter) minus the quantum efficiency before UV irradiation (QEbefore) divided by the quantum efficiency before UV irradiation (QEbefore) times 100%. The samples were irradiated in air. It was noticed that the Increase% varies from sample to sample while the passivating polymer remains the same. Especially, samples coated with PVB show large variations in the Increase%. Due to problems with the solubility of Na2S in ethanol,13 the reproducibility of synthesis of samples passivated with PVB was not very good. In the preceding report it has been shown that even small deviations in reaction conditions can cause large variations in luminescence properties.13 Nevertheless, a trend in the enhancement of the quantum efficiency is observed for samples coated with different polymers. Table 1 shows that samples passivated with PVB, MA, and PVA exhibit a substantial increase in quantum efficiency upon UV irradiation. It is well-known that when vinyl polymers such as PVB and PVA and the monomer MA are subjected to UV cross-linking reactions can occur (UV curing).26 Hence, the difference in Increase% of the luminescence intensity between the samples can be explained by the fact that PVB, PVA, and MA have a possibility to polymerize further and form crosslinks, while for PP no further polymerization can occur. A higher degree of polymerization can lead to a better coverage of the surface of the particles and can reduce the number of dangling bonds on the surface of the particle. Dangling bonds provide surface trap states for nonradiative recombination.21

Yet, if the UV enhancement of the luminescence is only due to the further cross-linking and polymerization of the passivating molecules at the surface of the nanoparticles, one would expect no UV-enhancement effects for ZnS:Mn2+ nanoparticles without a polymer coating, and samples coated with polymers that have no possibility to polymerize and form cross-links, like PP. As can be seen in Table 1, nanoparticles coated with PP and without passivating polymer are also sensitive to UV irradiation. Apparently, there is a second process that can cause an increase in quantum efficiency. In the literature several possibilities have been suggested. Dunstan and co-workers18 observed that upon UV illumination of nanocrystalline ZnS the emission bands at 420, 480-500, and 550 nm increased. These peaks were tentatively associated with respectively SO42- adsorbed at the surface, Zn at the surface, and S at the surface. These are all products of a photocorrosion process that was postulated to take place during the UV irradiation. Thus, upon irradiation of nanocrystalline ZnS, more radiative recombination centers should be formed. This would result in a higher quantum efficiency. However, it is not clear how SO42-, Zn, and S at the surface could serve as luminescent recombination centers, since these ions and atoms do not luminesce in the mentioned region (SO42-), or do not luminesce at all (Zn, S). Another mechanism was proposed by Henglein et al.17 They observed that oxygen influences the fluorescence of colloidal ZnS in a complex manner. The fluorescence intensity of ZnS colloids increased under prolonged illumination in the presence of oxygen. This was explained by photoanodic dissolution of nanocrystalline ZnS during the illumination. The particles become smaller, resulting in an increase in the fluorescence yield, which was also described for colloidal CdS.17 In the present case we are not dealing with colloidal solutions, so photoanodic dissolution cannot take place. However, one could think of some kind of photooxidation process taking place during the irradiation. Photooxidation products (e.g., ZnSO4) can block nonradiative pathways as was proposed by Becker and Bard.16 To investigate the role of oxygen and other gases in the UVenhancement process, the samples were irradiated in different atmospheres. 3.2. Influence of the Atmosphere on the UV Enhancement. To investigate the influence of the atmosphere on the UV enhancement, samples coated with MA and PP and the sample without passivating polymer were UV irradiated in pure nitrogen, dry air (20% O2, 80% N2), and humid nitrogen (N2 + H2O, exact composition not known). The Increase% of the quantum efficiency and the irradiation times needed to reach maximum luminescence intensity of the above-mentioned samples obtained in different atmospheres are shown in Table 2. 3.2.1. Irradiation in an Inert Atmosphere. To deduce whether air (containing nitrogen, oxygen, water, and small amounts of other gases) is playing a role in the UV-enhancement process, the samples were irradiated in an inert nitrogen atmosphere. Before irradiation the sample chamber was thoroughly flushed with nitrogen, to be sure that air was absent during the irradiation.

TABLE 2: Increase% of the Luminescence Intensity and Irradiation Time (tirr) Measured for Samples Coated with MA and PP and Sample without Passivating Polymer in Air, Pure Nitrogen, Dry Air, and Humid Nitrogen air

nitrogen

dry air

humid nitrogen

polymer

Increase%

tirr (min)

Increase%

tirr (min)

Increase%

tirr (min)

Increase%

tirr (min)

MA PP without

240 60 100

60 200 30

140 0 50

170 600 35

590 20 180

60 375 425

780 200 370

50 150 100

10206 J. Phys. Chem. B, Vol. 105, No. 42, 2001 As can be deduced from Table 2, the sample coated with MA shows a modest increase of the quantum efficiency in an inert atmosphere (140%) which is less than that in air (240%). The observed increase in luminescence intensity in nitrogen can be explained by UV curing. The UV radiation can initiate further cross-linking by free-radical polymerization. The lower Increase% in nitrogen can be caused by the absence of oxygen. Oxygen can promote the polymerization and cross-linking of MA.26 Besides, the absence of air possibly prevents the second process, which causes the UV enhancement of the PP-coated sample and the sample without passivating polymer. It is possible that both processes play a role in the UV-enhancement process of samples coated with MA. The sample passivated with PP shows no UV enhancement of the luminescence intensity in inert atmosphere. This means that air is indeed needed to cause UV enhancement for this sample. The sample without passivating polymer still shows some UV enhancement in N2. However, the Increase% in nitrogen is less than that in air. Perhaps, some water or oxygen is adsorbed at the surface of the nanoparticles, which cannot be removed during the flushing of the sample chamber with N2. This can cause the observed small Increase% in nitrogen. In view of the very low quantum efficiency of the unpassivated sample (0.3%), nonradiative decay dominates and a small amount of surface passivation can give an observable increase in the luminescence efficiency. From the experiments described above, it is clear that air plays a role in the UV enhancement. Air is a mixture of gases (78% N2, 21% O2, 1% Ar, 0.03% CO2, 0.01% other27). Furthermore, the average relative humidity of the air in The Netherlands is around 60%.28 It is known that both water and oxygen can react photochemically with the surface of the ZnS nanoparticles upon UV irradiation. Therefore, to distinguish between the role of oxygen and water in the UV-enhancement process, the samples were irradiated in both dry air and humid nitrogen atmospheres. 3.2.2. Irradiation in Dry Air. Dry air was obtained by bubbling air through concentrated sulfuric acid. From Table 2 it can be concluded that the amount of UV enhancement in dry air is dependent on the polymer which was used to passivate the surface of the nanoparticles. The sample coated with MA shows the highest Increase% (590%) followed by the sample without passivating polymer (180%). The sample passivated with PP shows only a small increase in quantum efficiency upon irradiation in dry air (20%). The irradiation time needed to reach maximum luminescence intensity is also highly dependent on the polymer used to passivate the nanoparticles. The sample coated with MA reached in 60 min its maximum intensity, while for the PP-coated sample and the sample without passivating polymer much longer irradiation times were needed (375 and 425 min, respectively). This supports the existence of two mechanisms for the observed UV enhancement. The UV curing of the MA-coated sample appears to be a very fast process. Within 1 h the quantum efficiency increased by a factor of 5.9. The second process, in which probably photooxidation of the surface plays a role, is a much slower process. It takes around 6.5 h to complete this process. Furthermore, photooxidation affects the quantum efficiency less than UV curing, since the MA-coated sample shows a much higher Increase% than the sample without passivating polymer (590% and 180%, respectively). The effect of the irradiation in dry air on the excitation spectrum of the Mn2+ emission was also investigated. In Figure 3 several excitation spectra are shown of nanocrystalline ZnS: Mn2+ without passivating polymer measured at different ir-

Bol and Meijerink

Figure 3. Exctiation spectra (λem ) 590 nm) of nanocrystalline ZnS: Mn2+ without passivating polymer measured after different irradiation times in dry air.

radiation times. The maximum as well as the onset of the excitation spectrum (recorded at λem ) 420 nm or λem ) 590 nm) shifts to shorter wavelengths during the UV irradiation. The samples with passivating polymer (MA, PP) show the same kind of behavior. Earlier, a shift of the maximum was observed by Gallagher et al.21 for particles coated with MA and PMMA. To explain this shift, they suggested that the smallest particles in the sample are most sensitive to UV curing because the proportion of surface sites is the largest in these particles. Therefore, the UV-curing process would enhance the relative contribution of the smallest particles to the overall luminescent efficiency, which would cause a shift in the maximum of the excitation spectrum to shorter wavelengths during the UV irradiation. However, this mechanism does not explain the blue shift of the onset in the excitation spectrum upon irradiation, as is observed here. Another mechanism that can explain the observed shift of the excitation spectrum is photoinduced oxidation of ZnS. In the presence of oxygen, the surface of the ZnS nanoparticles can be photooxidized to, e.g., ZnSO4, resulting in a decrease of the ZnS particle size, and a shift of the excitation spectrum (including the onset) to shorter wavelengths occurs due to quantum confinement effects. A similar procedure is used for making nanocrystalline Si particles.29 A heat treatment results in the reaction of the surface of Si particles to form SiO2. The decrease of particle size is monitored by a blue shift of the absorption maximum as a function of heating time. The increase of the quantum efficiency by photooxidation of the surface of the nanoparticles can be due to two effects. Possibly, smaller ZnS nanoparticles show higher quantum efficiency,17 although clear evidence for this effect has not been provided yet. Another possibility is that the photooxidation products such as ZnSO4 serve as a good passivating layer on the ZnS:Mn2+ particles and reduce nonradiative recombination at the surface. The onset of the excitation spectrum of the sample coated with MA also shifts to the blue upon irradiation in dry air. This suggests that besides UV curing, which cannot explain a shift of the onset of the excitation spectrum, photooxidation occurs as well. In conclusion, The ZnS nanoparticles become smaller upon irradiation in dry air due to photooxidation. Photooxidation products most probably passivate the surface of the PP-coated sample and the sample without passivating polymer, which results in higher quantum efficiency. For the sample coated with

Luminescence Quantum Efficiency of ZnS:Mn2+. 2 MA both UV curing and photooxidation play a role in the UVenhancement process in dry air. 3.2.3. Irradiation in Humid Nitrogen. A humid nitrogen atmosphere was created by bubbling nitrogen through water. The results of the irradiation in humid nitrogen are shown in Table 2. From Table 2 it can be deduced that the Increase% of the luminescence intensity is even higher in humid nitrogen than it is in dry air. For the PP-coated sample the Increase% obtained in humid nitrogen is 10 times higher than that in dry air (200% and 20%, respectively). The sample without passivating polymer exhibits the same trend (370% and 180%, respectively). The sample passivated with MA has also a higher Increase% (780%) in humid nitrogen. Possibly, the products of the photochemical reaction that takes place upon irradiation in the presence of water passivate the surface better than the photooxidation products obtained during irradiation in dry air. One of the products that could be created during the irradiation in the presence of water is Zn(OH)2. It is known that a Cd(OH)2 layer around CdS nanoparticles passivates the surface very well. The luminescence quantum efficiency increases from less than 1% to more than 50% by capping CdS nanoparticles with a hydroxide layer.30 In addition, the irradiation time needed to reach maximum luminescence intensity in humid nitrogen is shorter than it is in dry air, especially for the samples coated with PP and without passivating polymer. The photochemical reaction taking place in humid nitrogen is apparently faster than the photooxidation reaction, which takes place in the presence of oxygen. The excitation spectra as a function of the irradiation time obtained in humid N2 show the same behavior as in dry air: the excitation spectra shift to higher energy upon irradiation for all samples indicating that the particle size decreases upon irradiation in humid N2. In dry nitrogen a shift of the excitation spectrum was not observed upon irradiation. If the increase in quantum efficiency is caused by the decreasing particle size, as was proposed by Henglein,18 a larger shift of the excitation spectrum was expected in humid nitrogen compared to dry air, since the Increase% values in the humid N2 atmosphere are higher. However, the excitation spectra measured in humid nitrogen do not shift further to the blue than they do in dry air upon irradiation. This supports our interpretation that the increase of the luminescence quantum efficiency is due to the creation of a new passivation layer during the irradiation. The products formed during irradiation in the presence of water passivate the surface better than those formed in the presence of oxygen. 3.2.4. Photochemical Decomposition. By comparing the Increase% values obtained in air with those obtained in dry air and humid nitrogen, it can be concluded that in an environment with only oxygen or water the UV enhancement is stronger than in an environment where both gases are present. If samples are irradiated in air, the color of the samples without passivating polymer and the MA-coated samples changed from white to yellowish brown, which causes a drop in the luminescence intensity after the point that the intensity had reached its maximum. The UV-induced coloration can be explained by photochemical decomposition of the ZnS host. Photochemical decomposition of ZnS crystals or powders is a well-known phenomenon. Blackening of ZnS in sunlight was already observed in the 19th century when paintings with the white paint lithophone (mixture of BaSO4 and ZnS) colored gray when they were exposed to sunlight.31 This observation has led to numerous experiments (see, e.g., refs 31-33) to explain this behavior, to investigate how to avoid blackening,

J. Phys. Chem. B, Vol. 105, No. 42, 2001 10207

Figure 4. Typical time trace of the luminescence intensity at 590 nm of nanocrystalline ZnS:Mn2+ (PP coated) (λexc ) 300 nm). The dark periods are indicated in the figure.The first steep decrease of the luminescence intensity (after a dark period of 120 min) is enlarged in the inset.

and to make new photostable white paints. From these experiments it became clear that the blackening originates from the photochemical decomposition of the ZnS and proceeds via a series of photochemical reactions with Zn, S, and ZnSO4 as end products.31 Water enhances the blackening of the ZnS.32 The presently observed blackening and the decrease in the photoluminescence intensity after prolonged irradiation can be explained by the formation of these products at the surface of the ZnS:Mn2+ nanoparticles. These products prevent efficient absorption of the UV radiation used to excite the sample and absorb the emitted light, which both lead to a decrease of the quantum efficiency. Apparently, if both oxygen and water are present during the UV irradiation, the photodecompostion rate is higher than when only oxygen or water is available. This results in a smaller Increase% in air and, after prolonged irradiation, a decrease in luminescence intensity. The absence of coloration of samples passivated with PP indicates that PP prevents the photodecomposition of the ZnS. 3.3. UV-Induced Quenching. In the previous sections the increase in quantum efficiency of the Mn2+ emission due to UV irradiation is described. Depending on the passivating polymer and the atmosphere, the luminescence efficiency increases for UV-irradiation periods of minutes to hours. However, on a short time scale an opposite effect is observed. In the first seconds of UV irradiation a 30% drop of the initial intensity was observed. A typical experiment is shown in Figure 4. When UV irradiation starts, after a period of darkness the emission intensity falls rapidly, then more slowly to stabilize after about 5 min. After the signal has stabilized, an increase may be observed due to UV-enhancement effects if the sample has not been UV irradiated for a long time. A similar observation has been made before by Weller et al.34 for the defect-related emission of nanocrystalline ZnS particles in solution. A very plausible model was presented for these effects. The sudden drop in the emission intensity was ascribed to the UV-induced formation of O2-. Oxygen is able to capture a photogenerated electron from the conduction band and an O2- molecule is formed, which adsorbs at the ZnS surface and acts as an efficient quencher of the ZnS emission. The slow recovery of the luminescence signal in the dark was assigned to slow desorption of O2- molecules. In the present experiments the same effects are observed for the Mn2+ emission in nanocrystalline ZnS particles. In the dark the initial emission intensity slowly recovers. In Figure 4 it can be observed that the longer the dark

10208 J. Phys. Chem. B, Vol. 105, No. 42, 2001

Figure 5. Initial luminescence intensity at 590 nm (λexc ) 300 nm) of nanocrystalline ZnS:Mn2+ (PP coated) as a function of the dark period before UV excitation starts.

period is, the higher the initial emission intensity rises. In Figure 5 the initial emission intensity is plotted as a function of the dark period and it can be observed that a full recovery of the emission intensity is not observed until after dark periods of 1 h or more. This shows that the recovery, related to the disappearance of the quenching species, is a very slow process. To further investigate the origins of the initial sharp decreases in the emission intensity such as those in Figure 4, measurements were performed for many different ZnS:Mn2+ samples and in different atmospheres. Also the rate of recovery of the signal was measured as a function of temperature. The influence of the type of sample showed that the weakly emitting samples (such as ZnS:Mn2+ without passivating polymer) showed smaller drops (only a few percent) in the emission intensity. This can be explained qualitatively by considering that the fast drop is due to the formation of additional quenching centers. The extra quenching sites will have less influence for samples with a low quantum efficiency than for samples with high quantum efficiencies and relatively few quenching sites. To study the role of oxygen in the fast UV-induced quenching, experiments were performed in different atmospheres. In dry nitrogen also a drop in the emission intensity was observed, not much different from the results found in air. Although it might be argued that this is due to residual oxygen adsorbed on the surface of the nanoparticles, it indicates that O2- is not responsible for the UV-induced quenching. The repeated quenching and recovery ascribed to oxygen desorption in the model in ref 34 is not expected in a dry nitrogen atmosphere if some initially adsorbed oxygen is responsible. Also, in a He atmosphere measurements very similar to those shown in Figure 4 were measured. These experiments suggest that UV-induced quenching is related to another mechanism. In an atmosphere with ethanol (made by bubbling nitrogen through ethanol) no initial drop is observed and the emission is constant at the initial (high) level. Bubbling through water reduces the initial quenching. To study the influence of the temperature, the recovery of the signal for various dark periods between 30 and 150 s were measured for temperatures between 20 and 75 °C. No strong influence of temperature on the recovery is observed. If the recovery was due to O2- from the surface, a significant influence of temperature would be expected, since desorption is expected to be faster at higher temperatures. Based on the experiments described above, it can be concluded that under UV irradiation nonradiative recombination centers are formed within seconds. After a while a steady state

Bol and Meijerink situation is realized in which a constant number of (extra) quenching sites partly quench the Mn2+ (or ZnS) emission. In the dark the quenching centers slowly disappear. The fact that the effects are also observed in a dry nitrogen or helium atmosphere indicated that oxygen is not responsible for the fast UV-induced quenching, as has previously been suggested.34 The observation that the recovery rate is not enhanced by increasing the temperature between 20 and 75 °C is also in contradiction with the previous model.34 A similar fast decrease in luminescence intensity has been reported for phosphors such as Sr3Gd2Si6O18:Pb,Mn35 and Ca10(PO4)6(F, Cl)2:Sb,Mn.36 These phosphors show a fast degradation as a consequence of 185 nm irradiation. A 10-25% drop was observed within 6 min with a very steep decrease in the first minute. This phenomenon was called short-term degradation (STD) and was explained by the fast trapping of charge carriers at lattice defect states. These centers are quenching centers for luminescence. On the basis of the present experiments, it is not clear what UV-induced center causes the fast drop in emission intensity. The absence of a drop in an atmosphere with ethanol indicated that it may be related to a trapped hole center (ethanol is able to scavenge holes and thus removes these centers). The fact that the recovery rate is not influenced by temperature points to a rather deeply trapped hole center that acts as a nonradiative recombination site. For a more definite assignment of the nature of the site responsible for the quenching, EPR experiments could be performed. The slow disappearance of the quenching site can be related to change in the in the EPR spectrum in the dark, assuming that the trapped hole site or possibly another UVinduced quenching site is EPR-active. If the nature of the quenching site is identified, it may help to make more efficiently luminescing nanocrystalline ZnS:Mn2+ samples by adjusting the synthesis procedure to prevent the formation of these sites. 4. Conclusions UV enhancement of the luminescence quantum efficiency of nanocrystalline ZnS:Mn2+ is dependent on both the passivating polymer and the ambient atmosphere. Two mechanisms play a role: UV curing and a photochemical reaction process at the surface of the nanoparticles. The UV enhancement for samples coated with PVB, PVA, and MA is predominantly caused by the UV curing of the passivating polymer. UV-induced crosslinking results in a better passivating layer around the surface of the nanoparticles. The UV enhancement observed for PPcoated samples and the unpassivated samples is explained by photochemical reactions. The products of these reactions (e.g., ZnSO4 or Zn(OH)2) can serve as a passivating layer around nanoparticles, which increases the quantum efficiency. Water or oxygen is required for these photochemical reactions. After UV curing quantum efficiencies of about 10% are obtained. Prolonged UV irradiation (hours to days) in the presence of water and oxygen also leads to coloration and a decrease of the luminescence quantum efficiency. In addition to the long-term UV enhancement of the Mn2+ emission intensity, short-term (seconds) UV-induced quenching of the Mn2+ emission is observed. The short-term quenching is related to the UV-induced filling of hole traps and not to the formation of O2-, as was suggested previously. Acknowledgment. The financial support of Philips Lighting NV is gratefully acknowledged. We thank Ms Helen de Waard for performing the ICP analysis.

Luminescence Quantum Efficiency of ZnS:Mn2+. 2 References and Notes (1) Rossetti, R.; Hull, R.; Gibson, J. M.; Brus, L. E. J. Chem. Phys. 1985, 82, 552. (2) Brus, L. J. Phys. Chem. 1986, 90, 2555. (3) Henglein, A. Chem. ReV. 1989, 89, 1861. (4) Wang, Y.; Herron, N. J. Phys. Chem. 1991, 95, 525. (5) Alivisatos, A. P. J. Phys. Chem. 1996, 100, 13226. (6) Bhargava, R. N.; Gallagher, D. Phys. ReV. Lett. 1994, 72, 416. (7) Bhargava, R. N. J. Lumin. 1997, 72-74, 46. (8) Murase, N.; Jagannathan, R.; Kanematsu, Y.; Watanabe, M.; Kurita, A.; Hirata, K.; Yazawa T.; Kushida, T. J. Phys. Chem. B 1999, 103, 754. (9) Bol, A. A.; Meijerink, A. Phys. ReV. B 1998, 58, R15997. (10) Sooklal, K.; Cullum, B. S.; Angel, S. M.; Murphy, C. J. J. Phys. Chem. 1996, 100, 4551. (11) Koshravi, A. A.; Kundu, M.; Kuruvilla, B. A.; Shekhawat, G. S.; Gupta, R. P.; Sharma, A.; K. Vyas, P. D.; Kulkarni, S. K. Appl. Phys. Lett. 1995, 67, 2506. (12) Leeb, J.; Gebhardt, V.; Mu¨ller, G.; Haarer, D.; Su, D.; Giersig, M.; McMahon, G.; Spanhel, L. J. Phys. Chem. B 1999, 103, 7859. (13) Bol, A. A.; Meijerink, A. J. Phys. Chem. B 2001, 105, 10197. (14) Pingbo, X.; Weiping, Z.; Min, Y.; Houtong, C.; Weiwei, Z.; Liren, L.; Shangda, X. J. Colloid Interface Sci. 2000, 229, 534. (15) Isobe, T.; Igarashi, T.; Konishi, M.; Senna, M. Mater. Res. Symp. Proc. 1999, 536, 383. (16) Becker, W. G.; Bard, A. J. J. Phys. Chem. 1983, 87, 4888. (17) Henglein, A. Top. Curr. Chem. 1988, 143, 151. (18) Dunstan, D. E.; Hagfeldt, A.; Almgren, M.; Siegbahn, H. O. G.; Mukhtar, E. J. Phys. Chem. 1990, 94, 6797. (19) Yu, I.; Liu, H.; Wang, Y.; Fernandez, F. E.; Jia, W. J. Lumin. 1996, 76-77, 252.

J. Phys. Chem. B, Vol. 105, No. 42, 2001 10209 (20) Yu, I.; Liu, H.; Wang, Y.; Fernandez, F. E.; Jia, W.; Sun, L.; Jin, C.; Li, D.; Liu, J.; Huang, S. Opt. Lett. 1997, 22, 913. (21) Gallagher, D.; Heady, W. E.; Racz, J. M.; Bhargava, R. N. J. Mater. Res. 1995, 10, 870. (22) Yu, I.; Isobe, T.; Senna, M. J. Phys. Chem. Solids 1996, 57, 373. (23) Cullity, B. D. Elements of X-ray Diffraction; Addison-Wesley: Massachusetts, 1978; p 102. (24) Lu, S. W.; Lee, B. I.; Wang, Z. L.; Tong, W.; Wagner, B. K.; Park, W.; Summers, C. J. J. Lumin. 2000, 92, 73. (25) Oda, S.; Kukimoto, H. J. Lumin. 1979, 18-19, 829. (26) Stevens, M. P. Polymer chemistry; Addison-Wesley: Reading, MA, 1975. (27) Lide, D. R. Handbook of Chemistry and Physics, 74th ed.; CRC Press: Boca Raton, FL, 1993. (28) Source: KNMI (Royal Dutch Meteorologic Institute). (29) Brongerma, M. L.; Polman, A.; Min, K. S.; Boer, E.; Tambo, T.; Atwater, H. A. Appl. Phys. Lett. 1998, 72, 2577. (30) Spanhel, L.; Haase, M.; Weller, H.; Henglein, A. J. Am. Chem. Soc. 1987, 109, 5649. (31) Platz, H.; Schenk, P. W. Angew. Chem. 1936, 49, 822. (32) Lenard, P. Ann. Phys. 1922, 68, 553. (33) Schleede, A. Z. Phys. Chem. 1923, 106, 390. (34) Weller, H.; Koch, U.; Guttie´rrez, M.; Henglein, A. Ber. BunsenGes. Phys. Chem. 1984, 88, 649. (35) Verhaar, H. C. G.; Kemenade, W. M. P. van. Mater. Chem. Phys. 1992, 31, 213. (36) Flaherty, J. M. J. Electrochem. Soc. 1981, 128, 131.