1. Introduction

Persistent phosphor materials, having a long afterglow after the optical excitation has ceased, have been known since a long time. Recently, they have gained renewed interest as several newly developed materials show visible luminescence many hours after excitation, opening up lots of novel applications. One of the most important of these is power-less safety illumination in case of electric power failures [1]. The usual way in which these phosphors are evaluated, is by measuring their photometric output as a function of time. The afterglow intensity after -for example - 10 minutes is then quoted as a figure of merit. Alternatively, an extinction time is defined as the time until the luminance has decayed to 0.32 mcd/m², which is roughly 100 times the eye sensitivity in dark-adapted condition [2]. Both methods use the photometric quantity candela, which is defined in the context of high-intensity color vision, where only the cones in the human retina are assumed to contribute to vision. However, below about 1 cd/m², rods in the retina start to contribute to the perception of luminance, and below 1 mcd/m², rods have taken over completely. Rods have a much higher sensitivity than cones and their sensitivity curve is shifted towards the blue compared to cones. This results in the Purkinje effect, named after the 19th century Czech physiologist who observed that red objects which appeared brighter than blue or green objects in daylight, appeared darker than the blue and green objects at dawn [3]. Therefore, a new, unified luminance definition, as developed by Rea at al. [4], is used to more accurately describe the low-level performance of light sources such as persistent phosphors. We will see that this sheds an entirely new light on the relative performance of phosphors with different emission spectra. The effect is illustrated using a number of selected ‘benchmark’ commercial persistent phosphor powders.

2. Experimental

A set of four commercial persistent phosphor powders with the same particle size and different emission spectra, spanning the whole visible region, was selected as benchmark materials. Materials were purchased from GloTech International [5]. Properties of these benchmarks are shown in Table 1. The composition of the host materials was determined using X-ray diffraction (Siemens D5000, CuKα radiation). While these phosphors use a different crystal host, the emitting center is europium in all cases. Eu is built in substitutionally as Eu²⁺ for all but the red-emitting phosphor, where it is incorporated as Eu³⁺. Details on the luminescence properties of these materials can be found elsewhere [6, 7, 8, 9, 10]. Emission spectra were measured both upon continuous excitation and after excitation, using a calibrated Ocean Optics QE65000 fiber optic coupled CCD spectrometer. Absolute luminance of the afterglow was measured with the same equipment after calibration of the CCD spectrometer against a lux meter.

3. Results and discussion

Common measurements of perceived brightness of light sources take into account the human photopic sensitivity curve, which is peaking at 555 nm, with a maximum value of 683 lm/W (Fig. 1). At very low luminance, only the rods in the retina are active (scotopic vision), and the corresponding sensitivity curve is also shown in Fig. 1. At these low light levels, it is seen that (1) the eye sensitivity is much larger and (2) the eye becomes more blue-sensitive and less red-sensitive. The real trouble begins if we have to describe the luminance of sources in the so-called mesopic or intermediate region, where both rods and cones are active. While several research groups have proposed a photometric standard to span the entire scotopic-mesopic-photopic spectrum (reviewed in [4]), there is at present no internationally accepted standard for a unified system of photometry. These efforts are complicated by a number of facts:

Table 1. Properties of commercial persistent phosphor materials, selected as benchmarks. PLE = Photopic luminous efficacy; S/P = ratio of scotopic to photopic efficacy.

View Table

 

Fig. 1. Human eye sensitivity functions at low (scotopic, V′ (λ)) and high light intensity (photopic, V (λ)); data are taken from [11].

Download Full Size | PDF

It is clear that a lot of research, including clinical tests and neuroscience, is needed to fully understand all these effects. It is realized by researchers that it will never be possible to find a system of photometry that covers all aspects of human vision [4, 15]. Nevertheless, the currently developed simple models, seem to converge to a unified approach that is close to international acceptance. We will use the model by Rea et al. [4] to estimate the characteristics of light sources such as persistent phosphors in the mesopic range. In this model, a unified luminance L is defined, in units cd/m². The value of L is supposed to be a more realistic measure of the perceived luminance of an illuminated object or light source; for very low luminance values, it should converge to the metric for scotopic vision, while at high luminance, L-values should be equal to the commonly determined photopic luminance values. In Table 1, some of the optical properties of the benchmark phosphors are shown. Herein, S and P are the scotopic and photopic luminances respectively. The latter is the quantity (in cd/m²) typically measured with any calibrated luminance meter. S and P are connected to the spectral power distribution I(λ) of a light source by:

and the photopic luminous efficacy (PLE), which equals the average eye sensitivity (in lm/W) for the given spectrum, is calculated as:

For monochromatic light sources, the PLE reduces to the value of V(λ) at the given wavelength. In the model by Rea et al., the wavelength dependent eye sensitivity in the mesopic range is estimated as a weighted average of the photopic V (λ) and scotopic V′ (λ) curves. If a light source has photopic and scotopic luminances P and S, the unified luminance L is given by

The parameter X is zero for unified luminances below 1 mcd/m² (scotopic vision) and one above 0.6 cd/m² (photopic vision); in between these values, for mesopic vision, values of X are considered linearly dependent on unified luminance:

When this equation for X is substituted in (4), one obtains a quadratic equation which can easily be solved for L. From these equations, it follows that L is only dependent on P and S or, alternatively, on the photopic luminance P and the ratio S/P. The resulting unified luminance L is shown as a function of wavelength in Fig. 2 for a number of different values of photopic luminance. This figure is drawn for the case of monochromatic light sources at the given wavelengths.

 

Fig. 2. Unified luminance as a function of wavelength for different photopic luminance values. This figure is drawn for the case of monochromatic light sources at the given wavelengths.

Download Full Size | PDF

 

Fig. 3. Persistent emission spectra of the selected benchmark phosphors of Table 1.

Download Full Size | PDF

It is seen that longer wavelengths are perceived less bright than short wavelengths when the intensity is very low. This is due to the combined effect of the cones becoming less active and the rods - being sensitive to shorter wavelengths - becoming more active. At a wavelength of 555 nm, the photometric and unified luminances coincide, since at that wavelength, sensitivities of rods and cones are assumed identical (Fig. 1). In order to evaluate the effect of the eye sensitivity shift upon unified luminance in case of actual emission spectra, a series of 4 commercially available powder phosphors was selected. The most important properties of these are summarized in Table 1, and the normalized spectra of the persistent emission are shown in Fig. 3. From the latter figure, it is seen that the materials were chosen as to emit in different parts of the visible region. The S/P value, shown in Table 1, together with the photopic luminance P, is used as a figure of merit in Rea’s model: the parameter X in this model can be directly derived from P and S/P. As S/P is much higher than 1 for the violet emitting phosphor and much lower than 1 for the red phosphor, we expect these to show much higher and lower unified luminance than photopic luminance respectively. In Fig. 4, the measured persistent luminescence of all 4 benchmark phosphor materials is shown in the usual photometric units. Note that the afterglow intensity of all phosphors quickly falls below 1 cd/m², so that we are indeed in the mesopic range. As stated before, the quality of a persistent phosphor is specified as the luminance at a certain time or the time until the luminescence has decayed to a certain level, typically 0.32 mcd/m². It is seen that - according to these standards - the green phosphor is largely superior, and that the red and violet phosphors perform equally bad. The picture changes completely if we draw the same measurements on a graph with a unified luminance scale (Fig. 5). The short wavelength (violet) phosphor turns out to perform much better than the red one, its unified luminance being about ten times higher than its photopic luminance. The blue and green phosphors, which have a threefold difference in (photopic) luminance, show almost identical performance in terms of unified luminance (Fig. 5). Although we currently have no means to accurately evaluate the perceived brightness of these phosphors in a controlled environment, visual observation confirms the results from Fig. 5.

 

Fig. 4. Measured persistent luminescence in units of photopic luminance for the selected benchmark phosphors.

Download Full Size | PDF

 

Fig. 5. Measured persistent luminescence in units of unified luminance for the selected benchmark phosphors.

Download Full Size | PDF

4. Conclusions

The results presented in this paper are by no means the final answer to the description of luminance at all possible light levels, due to the complexities mentioned. However, the model used should provide a fairly accurate representation of how light levels in the mesopic range are perceived by the human eye. From the presented data, it is clear that the photometric quantities, as defined for photopic vision, are highly inaccurate at intermediate light levels. More specifically, low-wavelength emission is underestimated in brightness, while red emission is seriously overestimated. The latter effect compromises the development of red-emitting persistent phosphors. Indeed, due to the nature of human vision - shifting from cone to rod vision, with a corresponding change is absolute sensitivity and wavelength dependent sensitivity curve - low intensity red light is perceived much weaker than green or blue light with equal photometric luminance. In the near future, visual perception tests will be performed on the benchmark persistent phosphors used in this paper. This should allow us to quantitatively validate the methods presented here in practical situations. The unified luminance, proposed by Rea et al., is a worthy effort to quantify the observed brightness in the mesopic range. However, it is by no means obvious that the approximations made are fully applicable to the persistent phosphors in practical applications, such as electric power failures, as description of vision in the mesopic range is ‘not for the faint hearted’ [15].