4134
OPTICS LETTERS / Vol. 38, No. 20 / October 15, 2013
Effects of persistent luminescence decay on mechanoluminescence phenomena of SrAl2O4:Eu2+, Dy3+ materials Mohammad Reza Rahimi,1 Gun Jin Yun,1,* Gary L. Doll,1 and Jun-Seong Choi2 1
Department of Civil Engineering, The University of Akron, 244 Sumner St. ASEC 210, Akron, Ohio 44325-3905, USA 2
Korea Maintenance, Co. Ltd. KM Tower 103-5, Guro-5Dong, Guro-Gu, Seoul, South Korea *Corresponding author:
[email protected] Received July 3, 2013; revised September 9, 2013; accepted September 9, 2013; posted September 10, 2013 (Doc. ID 193235); published October 9, 2013
This Letter reveals for the first time, to the best of our knowledge, the effects of stress-free persistent luminescence (PL) decay on the mechanoluminescence (ML) phenomena and the effects of stresses and strain rates on the PL decay of SrAl2 O4 :Eu2 , Dy3 (SAOED) materials. Previous research on ML phenomena in this material has focused on the effects of strain rates and stress variations on ML light intensity. However, experimental evidence provided herein shows that the ML light emission is also related to the PL decay time elapsed until the onset of stressing and the PL decay rate is dependent on the stress, strain rate, and the stress-free PL decay time interval. For quantitative stress measurements using SAOED materials, understanding of ML light sensitivity and its dependence on critical factors (strain rate, stress-free PL decay time interval, photoexcitation time, instantaneous PL decay rate, etc.) is crucially important. This Letter provides new and important perspectives that are essential for developing predictive models and/or calibration procedures for ML stress sensors. © 2013 Optical Society of America OCIS codes: (280.4788) Optical sensing and sensors; (170.3650) Lifetime-based sensing; (260.3800) Luminescence; (310.4925) Other properties (stress, chemical, etc.). http://dx.doi.org/10.1364/OL.38.004134
During the past two decades, mechanoluminescence (ML) materials have been proposed as materials for smart sensors that can monitor structural damage and visualize stress distributions. Recently, they have shown promise for visualization of crack propagation [1], internal defects of pipes [2], and stress distribution in structures [3]. Optical fibers coated with triboluminescence (TL) materials have been embedded within fiber-reinforced composite materials for the purpose of internal damage detection. TL powder (i.e., ZnS:Mn) was also dispersed within cementitious materials for the purpose of damage monitoring [4]. Although all of these applications are based upon a linear relationship between ML light intensity and applied stresses, recent studies indicate that a possible nonlinearity may exist between the ML light intensity and the applied stress [5–7]. Rate-dependency of ML light emission is a well-known characteristic of ML sensing materials [8]. Recently, a lifetime-based stress measurement method has been proposed by Someya et al. [9]. A possibility of measuring stresses from rising and decaying time constants was proposed because time constants are not affected by concentrations and excitation times. However, the persistent luminescence (PL) decay time was not considered by simply subtracting the initial ML light intensity from the raw data. In order to use ML materials as quantitative nondestructive evaluation (NDE) sensors, further understandings of ML mechanisms are needed beyond the proportionality of ML light intensity to the applied stress and strain rate effects. In this Letter, ML phenomena of SrAl2 O4 :Eu2 , Dy3 (SAOED) thin film materials have been examined focusing on: (1) effects of PL decay time elapsed until the onset of loading, stress, and strain rates on ML light intensity, (2) effects of photoexcitation time on stressfree PL decay, and (3) effect of stress, strain rates, and stress-free PL decay time interval on instantaneous 0146-9592/13/204134-04$15.00/0
PL decay rate. Through interpretations of experimental observations within the context of a piezoelectrically induced detrapping model for the ML crystallites, we reassess ML light emission properties and draw critical conclusions that should form a foundation for future models of ML sensors. SAOED powders, a commercial phosphorescent material, were used in this study. Powders were mixed with an optical epoxy resin with a weight ratio of 1∶3. Then, thin films with approximate thicknesses of 0.02 in. (0.508 mm) were fabricated by the doctor-blade method. These films were secured onto dog-bonetype aluminum tensile specimens by a commercial adhesive. The ML film was photo-excited by a 40 W broadband light source (having wavelength range of 400–1000 nm) before conducting uniaxial tension tests. ML light intensity was measured from images captured by a CCD camera. While capturing images, ambient light was completely blocked to minimize its contribution to the ML light intensity. During testing, the films are under compressive stress in two directions (see Fig. 1), which promotes piezoelectrification within the SAOED crystallites. SAOED particles used in the tests have sizes of about 2–40 μm in diameter. Figure 2 shows SEM images of SAOED particles. Figure 3 shows ML intensity changes in response to loading applied at different stress-free PL decay time intervals (1 2, 3, and 4 min) and at different strain rates (0.1, 0.2, 0.3, and 0.4 mm∕s). In every experiment, the SAOED film was photoexcited for 2 min, and then loaded to the same force (15 kN). As expected, the ML light intensity was found to increase with increasing strain rate. However, an important observation from the graphs in Fig. 3 is the nonlinearity of the ML contribution to the light intensity. Additionally, the experimental results indicate that as the strain rate increases, the initial nonlinearity appears to be gradually diminished. © 2013 Optical Society of America
October 15, 2013 / Vol. 38, No. 20 / OPTICS LETTERS
4135
Fig. 1. Mechanism of ML phenomena of SAOED-based sensing films (ET, energy transfer; kT, thermal energy; V o , oxygen vacancy; V sr , strontium vacancy).
These data show that It is comprised of both PL and ML contributions. Although the mechanism behind the PL in these materials is controversial, we shall attempt to interpret these data using the PL model proposed by Aitasalo et al. [10] and for ML using the model of V. K. Chandra and B. P. Chandra [11]. These mechanisms are illustrated in Fig. 1. In the Aitasalo et al. model for PL, electrons excited in Eu2 luminescent centers escape into the conduction band and can become trapped at levels created by oxygen vacancies (V o ) and Dy3 ions. When sufficient thermal energy (kB T) is available, the trapped electrons can escape back into the conduction band and recombine in a luminescent center, and a ∼520 nm photon is emitted. Although the PL data appear to be linear in the region plotted in Fig. 3, PL decay curves of these materials are known to decay exponentially. V. K. Chandra and B. P. Chandra describe the mechanism for ML in these materials as piezoelectric fields arising from applied stresses created near activator ions decreasing the trap depths of the carriers (described above) or causing band-bending such that either thermal detrapping or tunneling to the conduction band may occur, respectively. The electrons moving in the conduction band can become captured in the excited state (4f 6 5d1 ) of the Eu ions that then emit a ∼520 nm photon as the ion relaxes to its ground state (4f 7 ). Based on these models, we therefore consider that whereas thermal energy causing detrapping of electrons is responsible for the PL, thermal energy plus piezoelectric fields are responsible for the ML. Another viewpoint of the effects of stress-free PL decay time interval on the ability to use the ML phenomena of SAOED in a sensor is shown in Fig. 4, where we
Fig. 2.
(a) and (b) SEM images of SAOED particles.
Fig. 3. Effects of stress-free PL decay time interval and strain rate on ML intensity (LI, light intensity).
plot the change of the relative ML light intensity, I peak − It0 0, where I peak is the peak light intensity at the load 15 kN and It0 0 is the light intensity at the time of loading for each combination of decay times (1, 2, 3, 5, and 7 min) and strain rates (0.1, 0.2, 0.3, and 0.4 mm∕s). Each datum is plotted as the average and standard error of the mean for three experiments and the solid lines are polynomial fits. Whereas the data in Fig. 4 show that the PL decay for t0 < 3 min is decreasing too rapidly to enable sufficient ML sensitivity to the lower strain rates of 0.1 and 0.2 mm∕s, there is adequate ML sensitivity for strain rates ≥0.3 mm∕s at t0 1 min. If this is a general trend, it implies that whereas impact stresses or fast crack propagation can be measured with high ML sensitivity in faster PL decay phase, slow static stresses can be measured with high ML sensitivity in slower PL decay phase (in this case, t0 > 3 min). Figure 5(a) shows that the stress-free PL decay in SAOED is dependent on the duration of the photoexcitation, a behavior that has been previously reported [12]. Stress-free PL decays of complex systems can be expressed as ItPL I 0 exp −t∕τb ;
(1)
Fig. 4. Effects of stress-free PL decay time interval and strain rate on relative ML light intensity, I peak − It0 0, where It0 0 is the light intensity at the time of loading for each combination of decay times (1, 2, 3, 5, and 7 min) and strain rates (0.1, 0.2, 0.3, and 0.4 mm∕s) (LI: light intensity).
4136
OPTICS LETTERS / Vol. 38, No. 20 / October 15, 2013
Fig. 5. Stress-free PL decay of SAOED depending on photoexcitation times.
where I 0 is the initial intensity, τ is the characteristic life, and b is the stretched exponent commonly observed in complex luminescence processes [13–15]. Arguments have been made that b is also a function of time [16]. In view of the PL model proposed by Aitasalo et al. [10], since the PL intensity must scale with the number of electrons decaying from the 4f 6 5d1 excited state of the Eu ion to the 4f 7 ground state, the increase in the overall intensity of the PL decay curves with photoexcitation time is consistent with increasing the population of electrons in the excited state. This qualitative interpretation implies that I o in Eq. (1) should approach saturation with photoexcitation time since the number of electrons that can take part in the PL process is finite. Consistent with this view, a plot of the areas under the PL decay curves shown in Fig. 5(b) shows that these numbers asymptotically approach a finite value. Because ML light emission of SAOED is strongly related to the PL decay characteristics, it is critical to have knowledge of the long-lasting PL decay of this material when it is incorporated into a sensor. Also plotted in Fig. 5(b) is the reduction in PL intensity at 300 s, a slowly changing region of the PL decay curve. For ML stress sensor applications, the temporal nature of the PL decay must be considered since it so significantly affects the relationship between the applied stress and the ML light intensity. In the foregoing, contributions of stress-free natural PL decay effects on ML phenomena were considered. However, our experimental test results showed that PL decay rates under static stresses are significantly different from the stress-free PL decay rates. Because stresses instantaneously change while loading increases, it is named instantaneous PL decay. Therefore, in order to investigate effects of stress, strain rates, and stress-free PL decay time interval on the instantaneous PL decay rate, a series of tests were conducted at specific decay times (1, 2, 3, and 5 min), and strain rates (0.1, 0.2, and 0.3 mm∕s). Loads were linearly applied up to different peak values (3, 6, 9, 12, and 15 kN) and maintained at the peak loads thereafter. Instantaneous PL decay trends are shown in Figs. 6(a)–6(d) for decay times of 1, 2, 3, and 5 min, respectively, at a strain rate of 0.3 mm∕s. Although results at other strain rates are not presented because of limited space, there are several important observations that can be made from the data. The PL decay rate following the rise to the peak intensity appears to be different from the stress-free PL decay curve corresponding to the same light intensity. Specifically, the PL decay
Fig. 6. Instantaneous PL decay under different static stresses at strain rate 0.3 mm∕s: (a) 1 min decay time, (b) 2 min decay time, (c) 3 min decay time, and (d) 5 min decay time, where the dashed blue line indicates stress-free PL decay.
becomes faster as stress increases. As decay times increase, the PL decay rate becomes slower. These observations are consistent with the view of increasing stresses and strain rates reducing the trap depth of oxygen vacancies and Dy3 ions, resulting in an increase of carriers escaping to the conduction band that then flood into the Eu2 luminescence centers and decay to the ground state through the emission of a photon. The presence of piezoelectric fields in SAOED results in faster decay rates than PL arising solely from thermal detrapping. From Eq. (1), the stretched exponent form of the luminescence intensity, the parameter b may be increasing with increasing piezoelectric field strength since it has been associated with the density of traps and trap release rates [17]. Finally, as decay time intervals increase, a decreasing b parameter associated with reduced trap release rates would cause PL decay rates under static stress to also decrease. Of course, since b is strongly influenced by the random distribution of site energies and intersite spacing [17], we expect that ML and PL characteristics of SAOED to be sample-dependent. V. K. Chandra and B. P. Chandra derived an expression for the rise in ML intensity It0 ML of materials like SAOED in their elastic regime as a function of applied stress (σ) and strain rate (_ε), where t0 is the timeframe measured relative to the onset of loading. In our experiments, load is applied linearly with time at a fixed rate so σ Eε Eαt0 and σ_ Eα. Their expression for It0 takes the form of _ − C 2 σ 2 C 1 E 2 α2 t0 − C 2 Eαt02 ; It0 ML C 1 σσ
(2)
where C 1 and C 2 are constants containing material parameters and E is the Young’s modulus. Provided that the stress-free PL decay (I PL ) and the instantaneous PL decay under static stresses (I 0ML ) are quantified by models in addition to predictable ML light intensity [I ML such as Eq. (2)], the experimentally measured intensity I exp shown in Fig. 3 will be predictable as
October 15, 2013 / Vol. 38, No. 20 / OPTICS LETTERS
I exp I 0 ; σ; ε_ I t I 0 ; σ; ε_ Z Z t t − I_ PL τdτ − I_ 0ML I 0 ; τ; ε_ dτ; td
(3)
td
where I_ is the rate of light intensity, I t indicates the total light intensity, which is proportional to the applied mechanical strain energy, and td is the time corresponding to I 0 . Examination of I exp plotted in Fig. 3 in view of the expression in Eq. (3) indicates several things. The I PL and I 0ML contribution to the curves is nonlinear. The stress-free PL decay rate corresponds to current I exp . The magnitude of I exp increases with strain rate, which is also consistent with increasing values of σ in Eq. (2). Development of models for I t , I PL . and I 0ML requires comprehensive experimental data. However, details of the predictive models are not addressed in this Letter since they are not the focus herein. In this Letter, ML phenomena were investigated considering the effects of the PL decay time on the ML light intensity in addition to stress and strain rate effects. This new investigation is based upon experimental testing of SAOED thin films applied on aluminum specimens under monotonically increasing tensile forces. First, experimental test data provided herein showed that the ML sensitivity gradually decreases as the stress-free PL decay time interval increases at the fastest strain rate (0.4 mm∕s). However, as the strain rate decreases, the dependency of ML sensitivity on the PL decay time was reversed. Second, the stress-free PL decay (i.e., total PL decay and areas under PL decay curves) was characterized in terms of photoexcitation times. Third, this Letter reveals that PL decay rates under stress states are very different from stress-free PL decay rate corresponding to the same light intensity. This experimental evidence is explained through physical models associated with PL and ML phenomena. The PL and ML phenomena are determined by carriers’ lifetime and finite population and correlated to each other under the same emission mechanism. For quantitative stress measurements using ML materials, understanding of ML light sensitivity and its dependence on critical factors (strain
4137
rate, stress-free PL decay time interval, photoexcitation time, instantaneous PL decay rate, etc.) is crucially important. Findings from this investigation are crucial in the development of intensity-based predictive transduction or calibration models for ML sensing films. This research was possible with support from the Ministry of Knowledge Economy (South Korea) through 2012 development of industry root technologies (IT-fusion). The authors are grateful for their support. References 1. N. Terasaki, C. Li, L. Zhang, and C. N. Xu, in Sensors Applications Symposium (SAS) (IEEE, 2012), pp. 1–5. 2. D. Ono, C. N. Xu, C. Li, and N. Bu, J. Jpn. Soc. Exp. Mech. 10, 152 (2010). 3. W. X. Wang, T. Matsubara, Y. Takao, Y. Imai, and C. N. Xu, Mater. Sci. Forum 614, 169 (2009). 4. T. J. Dickens and O. I. Okoli, J. Reinf. Plast. Compos. 30, 1869 (2011). 5. G. J. Yun, M. R. Rahimi, A. H. Gandomi, G. C. Lim, and J. S. Choi, Smart Mater. Struc. 22, 055006 (2013). 6. C. N. Xu, in International Conference on Optical MEMS and Nanophotonics (2010), pp. 31–32. 7. M. Akiyama, K. Nishikubo, and K. Nonaka, Appl. Phys. Lett. 83, 650 (2003). 8. J. S. Kim, K. Kibble, Y. N. Kwon, and K. S. Sohn, Opt. Lett. 34, 1915 (2009). 9. S. Someya, K. Ishii, M. Saeki, and T. Munakata, Opt. Lett. 38, 1095 (2013). 10. T. Aitasalo, J. Holsa, H. Jungner, J. C. Krupa, M. Lastusaari, J. Legendziewicz, and J. Niitykoski, Radiat. Meas. 38, 727 (2004). 11. V. K. Chandra and B. P. Chandra, J. Lumin. 132, 858 (2012). 12. C. Pereyda-Pierre, R. Melendrez, R. Garcia, M. PedrozaMontero, and M. Barboza-Flores, Radiat. Meas. 46, 1417 (2011). 13. I. Fatkullin, K. Kladko, I. Mitkov, and A. R. Bishop, Phys. Rev. E 63, 067102 (2001). 14. B. M. Weon and J. H. Je, J. Appl. Phys. 97, 036101 (2005). 15. B. M. Weon and J. H. Je, Appl. Surf. Sci. 251, 59 (2005). 16. B. M. Weon, S. Y. Kim, J. L. Lee, and J. H. Je, Appl. Phys. Lett. 88, 013503 (2006). 17. L. Pavesi and M. Ceschini, Phys. Rev. B 48, 17625 (1993).
