Phosphorescent OLEDs: Lighting the Way for Energy-Efficient Solid-State Light Sources
Phosphorescent-OLED lighting is an emerging technology that offers power-efficient and high-quality illumination with compelling form factors such as thinness and flexibility. This article focuses on the development of OLED lighting panels, where phosphorescent emitters are used to realize high energy efficiency and long operational lifetime.
by Peter A. Levermore, Michael S. Weaver, Mike Hack, and Julie J. Brown
THE LIGHTING INDUSTRY is currently in transition. The incandescent lamp is still a principal source of illumination, despite its very low efficiency (e.g., 12 lm/W), and consumes a considerable portion of the world's electricity. New sources of energy-efficient lighting are therefore critical to the future reduction of worldwide energy consumption. Both light-emitting diodes (LEDs) and organic LEDs (OLEDs) have the potential to become new solid-state forms of general illumination, replacing current technologies with safe and energy-efficient alternatives. OLEDs offer a thin, lightweight, energy-efficient, and large-area diffuse source of lighting with excellent visual quality. Importantly, OLED lighting panels do not contain materials known to be hazardous. In addition to the environmental benefits, there is an aesthetic design dimension to OLED lighting that is not possible to replicate with fluorescent lamps or LEDs.
In recent years, the OLED lighting industry has undergone a period of rapid expansion. There have been frequent reports of high-efficacy laboratory test pixels1,2 and numerous demonstrations of large-area prototypes that explore the unique architectural potential of OLED lighting.3-7 In this article, we focus on the development of 15 x 15-cm phos-phorescent OLED (PHOLED) lighting panels that match efficiency (50 lm/W) with visual impact (total panel thickness of less than 2 mm).1 By using phosphorescent emitters, we demonstrate that it is possible to deliver energy-efficient illumination while maintaining the unique and attractive form factor of OLED lighting.
The basic principle of OLED operation is that electrons and holes are injected into organic films, where they combine to form excitons, which then generate light. The exciton can have a total spin of S = 0 (singlet state) or S = 1 (triplet state). Approximately 25% of generated excitons are thought to be in the singlet state, while 75% are in the triplet state. In terms of spin conservation and how it applies to photon emission, in fluorescent OLEDs only singlet excitons produce optical emission. Professors Stephen Forrest and Mark Thompson from Princeton University and the University of Southern California, respectively, first reported a major breakthrough in device efficiency based on phosphorescent emitters in OLEDs in 1998.8 Phosphorescent emitters contain a heavy metal atom that facilitates mixing of singlet and triplet states, allowing singlet-to-triplet energy transfer through intersystem crossing. Mixing of singlet and triplet states enables triplet states to radiate. Therefore, in phosphorescent devices, up to 100% of excitons can potentially produce optical emission, compared to approximately 25% in conventional fluorescent devices. This pioneering work by Forrest and Thompson, followed by the continuing development of phosphorescent OLEDs, is a critical technology that enables OLEDs to become an efficient and viable general illumination source.
Low-Cost Device Architecture
Several OLED device architectures can be used to achieve white emission: (a) multiple emitters in a single emissive region, (b) stacked OLEDs (SOLEDs) with multiple emissive regions, or (c) patterned monochrome OLEDs with an additional low-cost color-mixing layer. Here, we focus on multiple emitters forming a single emissive region, which is expected to be the architecture with the lowest manufacturing cost. For example, in an all-phosphorescent device, just six organic layers can be used: a hole-injection layer (HIL); a hole-transport layer (HTL); a red–green phosphorescent emissive layer (RG EML); an adjacent blue phosphorescent emissive layer (B EML); a blocking layer, one function of which can be to block charge migration; and an electron-transport layer (ETL). The inset of Fig. 3 shows a typical white-PHOLED architecture.
To reduce power consumption and extend device lifetime, a highly stable blue phosphorescent emitter is required. Here, we focus on the use of a light-blue emitter with a peak wavelength of 474 nm, 1931 CIE (x, y) coordinates of (0.17, 0.37), an external quantum efficiency (EQE) greater than 20%, and a luminous efficiency greater than 45 cd/A at 1000 cd/m2. This emitter is ideally suited to high-efficacy warm-white emission. For example, alongside phosphorescent red–green, it is straightforward to demonstrate high-efficacy white emission with a color-rendering index (CRI) of greater than 80 and a correlated color temperature (CCT) from 2700 to 4000 K.9,10
In addition to higher efficacy and reduced power consumption, the use of a phosphorescent blue emitter also simplifies the manufacturing process. The explanation of this is simple – when fluorescent blue is deposited alongside phosphorescent red–green, there is typically energy loss between adjacent layers. Specifically, the low triplet energy of the fluorescent blue emitter quenches emission from the phosphorescent material. Spacing layers are therefore required in hybrid phosphorescent–fluorescent white OLEDs, which adds to the manufacturing cost. In contrast, in the all-phosphorescent architecture shown in the inset of Fig. 3, spacing layers are not required. Manufacturing costs are therefore expected to be lower for this architecture than for SOLEDs (fewer deposition steps are required).
Phosphorescent OLED Lighting
Figure 1 shows a pair of 15 x 15-cm PHOLED lighting panels designed and fabricated at Universal Display Corporation (UDC) using the simple all-phosphorescent device architecture shown in the inset of Fig. 3.
Fig. 1: This pair of 15 x 15 cm (6 x 6 in.) PHOLED lighting panels was designed and fabricated at UDC. Panel 1, on the left, is designed for mainstream commercial lighting applications. Panel 2, on the right, is a design concept, where decorative bus lines define a lighting flower.
Panel 1 on the left of the photo is designed taking into considering mainstream commercial lighting applications,1 while Panel 2 on the right is a design concept, with decorative bus lines defining a lighting flower. A low-cost light-extraction film with a thickness less than 0.5 mm is used in each case to deliver 1.5x efficacy enhancement and realize uniform emission color across all viewing angles. The total thickness of each panel, including substrate, encapsulation glass, and light extraction film is less than 2.0 mm. These phosphorescent panels showcase the attractive thin form factor of OLED lighting and deliver high-quality light with extremely low power consumption.
In this article, we focus on Panel 1, the performance of which is summarized in Table 1. The emissive area is divided into equally sized squares, and bus lines are used to transport charge from electrode contacts located at the edge of the panel. The high-conductivity bus lines minimize resistive losses that would other-wise arise from the relatively low conductivity of the transparent conductive oxide (TCO) anode typically used for bottom-emission OLEDs. The panel drive voltage then approximately matches the equivalent pixel voltage, thereby maximizing panel efficacy from a voltage perspective. As a result, resistive heat-ing across the panel is minimized, which means panel temperature remains low, enabling excellent operational lifetime. The bus lines also provide the benefit of improved luminance uniformity, which ensures that any aging occurs uniformly across the panel. Spectrometer measurements confirm a luminance uniformity of 92% across the 15 x 15-cm panel after life-testing to LT70 (70% of the initial luminance).
When characterizing small-area OLED pixels, it is appropriate to quote luminance in units of cd/m2. However, when scaling up from pixels to large-area OLED lighting panels, it is also important to account for fill factor and total light output. Here, the critical parameter is luminous emittance in units of lm/m2, which expresses the total light output delivered by per unit area by the panel as perceived by the human eye. For an approximately Lambertian emitter, the conversion is simply luminous emittance (lm/m2) = π x luminance (cd/m2) x fill factor. As a general guide, for mainstream commercial lighting applications, a luminous emittance of approximately 9000 lm/m2 is required.11 For example, a typical fluorescent ceiling luminaire housing with three linear T8 fluorescent tubes has a fixture area of approximately 1 m2 and delivers a total output of approximately 9000 lm. An OLED lighting panel with a 70% fill factor must operate at a luminance of 4000 cd/m2 to deliver a luminous emittance of 9000 lm/m2. For OLED lighting to become a competitive general-lighting illumination source, it is essential that high power efficacy and operational lifetime are maintained at this high-luminance level. For other applications, a lower luminance may be used.
Figure 2 shows the efficacy (lm/W) and luminous efficiency (cd/A) of Panel 1 as a function of luminance (including 1.5x light-extraction-efficacy enhancement).
Fig. 2: Power efficacy (filled circles) and luminous efficiency (empty squares) versus luminance for Panel 1 are shown above. At 1000 cd/m2 (approximately 2200 lm/m2), Panel 1 has an efficacy of 50 lm/W. At 4000 cd/m2 (approximately 9000 lm/m2), Panel 1 has an efficacy of 37 lm/W.
At 1000 cd/m2 (approximately 2200 lm/m2), the efficacy of Panel 1 is 50 lm/W, while at 4000 cd/m2 (approximately 9000 lm/m2) the efficacy is reduced to 37 lm/W. Over the same range, the luminous efficiency falls from 64 to 59 cd/A. The observed slight drop in efficacy can be attributed to a small rise in drive voltage at higher luminance. The efficacy of Panel 1 is comparable to typical fluorescent luminaires (less than 50 lm/W when ballast and fixture losses are included).12 This exceptional OLED performance is achieved using all-phosphorescent emitters in a low-voltage architecture. In a later section, we propose a roadmap that describes a path forward from 50 to 150-lm/W white-PHOLED lighting panels.
An important consideration for OLED lighting panels is not only efficacy, but also the color and quality of light delivered. Standard metrics used to describe the color of light are CIE 1931 (x, y) coordinates and correlated color temperature (CCT). Figure 3 shows a plot adapted from Energy Star Program Requirements for Solid-State Lighting – Version 1.0.13
Fig. 3: Chromaticity and color temperature of Panel 1 (red star) and Pixel A (a UDC-made device referenced later on) are plotted against the Planckian curve. Inset shows the simple six-organic-layer device structure.
The Planckian blackbody curve is shown as a line passing through CIE 1931 (x, y) chromaticity space. Quadrangles are used to identify CCTs along the curve from 2700 to 7000 K. Each quadrangle defines the chromaticity range that is acceptable by Energy Star standards for a light source at each color temperature. For example, chromaticity coordinates too far above the quadrangles are considered too green, while coordinates too far to the left are considered too blue, etc. Quadrangle dimensions are based on seven-step MacAdam ellipses at each color temperature. Panel 1 has a CCT of 3000 K, with a CIE 1931 (x, y) coordinates of (0.447, 0.425). In this case, the CIE y co-ordinate is fractionally too high, although this could easily be corrected in future device optimization.
In CIE 1931 (x, y) color space, the MacAdam ellipse size varies with color temperature, dependent on the photopic response of the human eye. In order to compare differences in color, it is therefore instructive to convert into CIE 1976 (u¢, v¢) color space, where coordinate differences are proportional to perceived color differences. The conversion is very simple: u¢ = 4x / (–2x + 12y + 3) andv¢ = 9y / (–2x + 12y + 3). A measure known as duv (or Δuv) = (Δu2 + Δv2)1/2 can then be used to quantify how far the chromaticity of a light source lies from the blackbody curve. As a general rule, when designing an OLED lighting panel, one should target a Δuv < 0.005 with a CCT from 2700 to 7000 K.13 The chromaticity will then fall within one of the quadrangles shown in Fig. 3.
Of equal importance to the color of a light source is how well other colors are rendered by that light source. Although metrics such as the color-quality scale (CQS) have been developed in recent years,14 at present the only universally accepted measure of lighting quality is the color-rendering index (CRI). Standard test samples are used, and the CRI is rated on a scale of 0–100 (although negative CRIs are possible), with 100 meaning that all samples illuminated by the light source appear to standard observers to have the same color as when illuminated by a standard reference source. For color temperatures of 2000–5000 K, a blackbody radiator is used as the reference light source, while above 5000 K the reference is an agreed upon form of daylight.15 Typically, eight standard test samples (R1–R8) of low-to-medium saturation are used to calculate CRI, and this is the number that is quoted in most publications. Additional test samples (R9–R15) can also be included to calculate special CRIs. In particular, for certain light sources, a high R9 value is desirable, as this certifies effective rendering of deep red.
One of the innate advantages of OLED lighting is the broad emission spectra of organic molecules, which enable high-quality rendering of a wide range of colors. Unlike fluorescent lighting and inorganic LED lighting, a high CRI can therefore be achieved by OLED lighting without compromising efficacy. For example, using a phosphorescent light-blue emitter, Panel 1 has a CRI of 84 averaged across all viewing angles. Using a slightly more saturated blue emitter, a CRI > 90 is readily achievable for OLEDs.16 This quality of light emission is comfortably in excess of Energy Star criteria (CRI > 75) and is thought to be appropriate for mainstream indoor lighting fixtures (CRI > 80). More importantly, for Panel 1, there is also remarkably little variation in chromaticity or color rendering as a function of viewing angle (Δuv from 0 to 60° is 0.002). This exceptional uniformity with viewing angle is achieved using a low-cost and thin-form-factor light-extraction film that also delivers 1.5x efficacy enhancement.
The final consideration in OLED lighting panel design is operational lifetime. Here, it is imperative to design for low temperature to extend the lifetime of the organic materials. One critical element in reducing panel temperature is the use of phosphorescent red, green, and blue emitters, all with very high internal quantum efficiency (IQE). Minimal heat is then generated from non-emissive exciton states, ensuring significantly lower temperature and longer lifetime than an equivalent fluorescent OLED lighting panel. Table 1 shows that the lifetime of Panel 1 is LT70 >> 10,000 hours at 1000 cd/m2 (approximately 2200 lm/m2), with LT70 >> 1600 hours expected at 3000 cd/m2 (approximately 6750 lm/m2), and LT70 >> 1000 hours expected at 4000 cd/m2 (approximately 9000 lm/m2). This lifetime is already sufficient for initial niche lighting products where lower luminance is required. Importantly, there is also minimal shift in color with aging for Panel 1 with Δuv = 0.007 after aging to LT74 (74% of initial luminance). Normalized electroluminescence spectra before and after aging are shown in Fig. 4.
Fig. 4: Normalized electroluminescence (EL) intensity of Panel 1 is measured initially (red line) and after aging to LT74 (74% of initial luminance). There is minimal color shift with aging.
This result is extremely encouraging and demonstrates the exceptional stability assured by the simple all-phosphorescent OLED stack.
The Road Ahead
Phosphorescent OLED lighting efficacies are already comparable to fluorescent lighting efficacies, when one takes into account system-level losses, e.g., ballasts and optics. However, further improvements in efficacy and lifetime are essential. Figure 5 shows a roadmap of efficacy improvement from a current UDC 2010 status of 50 lm/W to about 150 lm/W for future OLED lighting panels by about 2020.11 About 160 lm/W appears to be the potential physical limit for the efficiency of OLED lighting panels.17
Fig. 5: A road map from 50 lm/W demonstrated by UDC in 2010 shows an increase to 150-lm/W OLED lighting in the future. Efficacy can be increased by lowering voltage, increasing IQE through improved charge balance, and developing low-cost and thin-form-factor light-extraction techniques.
Key areas where efficacy gains can be made are (a) reduced voltage through the development of lower-voltage EML and transport materials, (b) higher IQE through improved charge balance and the ongoing development of phosphorescent emitters that maintain efficiency at high luminance, and (c) techniques that extract light that would otherwise remain trapped inside the OLED device layers.17 The most significant efficacy gains are to be made through improved light extraction. For example, if outcoupling enhancement could be doubled from 1.5x (Panel 1) to 3.0x, then the efficacy could also be doubled. At higher efficacy, less heat would be generated and panel lifetime would also be improved.
An example that shows that 3.0x outcoup-ling enhancement is possible is a 113-lm/W white-PHOLED pixel with a CRI = 80 and CIE 1931 (x, y) coordinates of (0.441, 0.414), reported by UDC at SID 2010.1 This device is plotted as Pixel A in Fig. 3. In this instance, a high-index glass substrate was used to remove the optical barrier at the glass/anode interface, and an index-matching hemisphere macro-extractor was used to ensure all light rays propagate normal to the surface, ensuring maximum light extraction. The challenge for the future is to demonstrate the same order of light-extraction enhancement using outcoup-ling techniques that are both low cost and do not add thickness to the OLED lighting panel. This idea is shown schematically in Fig. 6, where light-extraction enhancement is plotted against added thickness for various outcoupling systems.
Fig. 6: The figure shows light-extraction enhancement as a function of thickness added. At present, the highest OLED efficacy is achieved using outcoupling techniques that add substantial thickness to the panel. The ultimate aim is to demonstrate a light-extraction enhancement of greater than 3.0x while maintaining an attractive thin form factor.
Unique Appeal of OLED Lighting
Phosphorescent-OLED lighting panels can now be engineered to produce power-efficient high-quality white light. However, the same could be said of other, more mature technologies, such as inorganic LED lighting. So what sets OLEDs apart from competing energy-efficient light sources? One straightforward answer lies in the revolutionary thin form factor of OLED lighting. OLEDs by nature produce diffuse light distributed over a large surface area that provides a refreshing and compelling alternative to point-source lighting. Operating at 4000 cd/m2 (approximately equivalent to 9000 lm/m2 assuming a fill factor of 70%), OLED panels can provide uniform, bright, and excellent visual quality illumination with very little glare. They can be viewed directly and admired for their simplicity without the added cost and complexity of baffles or louvers to mask the lighting element. In addition, OLED lighting panels can be transparent. All the organic layers in a PHOLED stack are transparent, so the use of transparent electrodes allows for a light source that is transparent in the off-state and can emit light through both surfaces when energized. By making OLED lighting panels consisting of individually addressable red, green, and blue stripes, it is also possible to make fully color-tunable OLED lighting panels with uniform appearance. Furthermore, these panels can be fully dimmable, offering rich luminance and color dynamics.
The UDC phosphorescent-OLED lighting prototypes in Figs. 7–918 showcase OLED panels that are thin, lightweight, transparent, and flexible, opening up exciting new applications and design concepts. Figure 7 shows a thin-form-factor desk lamp using Panel 1 as the illumination source. Figure 8 shows Transparent Light Origami (TLO), where red, green, and blue emission from transparent OLED panels is added to produce secondary colors and white light. Figure 9 shows a 15 x 15-cm lighting panel fabricated on a flexible metal foil substrate.
Fig. 7: This white PHOLED desk lamp was designed by Emory Krall, a designer at UDC. The delicate form is achieved using Panel 1, which provides an even and diffuse light source with a thickness less than 2 mm. The head can rotate a full 360° on carbon-fiber supports. The panel is fully dimmable.
Fig. 8: Transparent Light Origami (TLO) designed by Emory Krall, UDC, features transparent primary-color PHOLED lighting triangles mounted on adjustable hinges. When panels overlap, secondary colors of light appear. For example, red + green = yellow. White light appears when red, green, and blue panels all overlap along the line of sight.
Fig. 9: A flexible white-PHOLED panel designed and fabricated at UDC uses a 15 x 15-cm metal foil substrate held in a flexed arrangement. A top-emission device architecture is used with a transparent cathode and thin-film encapsulation.
The exceptional OLED characteristics provide an innovative design platform allowing previously unrealized integration of lighting and architecture.19 It is this unique marriage of exciting form factors, novel applications, and energy efficiency that ensures a bright future for phosphorescent-OLED lighting.
This work was supported in part by U.S. Department of Energy (DE-FC26-08NT01585, DE-EE0003253, DE-FG02-08ER85082, DE-FG02-07ER84810, DE-EE0000626, DE-FG02-07ER84809. and DE-SC0002122).
1P. A. Levermore, V. Adamovich, K. Rajan, W. Yeager, C. Lin, S. Xia, G. S. Kottas, M. S. Weaver, R. C. Kwong, R. Ma, M. Hack, J. J. Brown, SID Symposium Digest 41, 786 (2010).
2S. Reineke, F. Lidner, G. Schwartz, N. Seidler, K. Walzer, B. Lussem, and K. Leo, Nature 459, 234 (2009).
4GE http://www.ge.com/audio_video/ge/ innovation/ge_oleds_the_future_of_lighting.html.
6Philips (Lumiblade) https://www.lumibladeshop.com/index.php/.
7Osram (ORBEOS) http://www.osram-os. com/appsos/ORBEOS/index.html.
8M. A. Baldo, D. F. O'Brien, Y. You, A. Shoustikov, S. Sibley, M. E. Thompson, and S. R. Forrest, Nature 395, 151 (1998).
9S. Xia, P. A. Levermore, V. Adamovich, C. Lin, R. C. Kwong, M. S. Weaver, and J. J. Brown, MRS Proc. 1212, S2.2 (2010).
10P. A. Levermore, V. Adamovich, K. Rajan, W. Yeager, C. Lin, S. Xia, M. S. Weaver, R. Kwong, R. Ma, M. Hack, and J. J. Brown,Proc. SPIE 7617, 761704 (2010).
11U.S. Department of Energy Solid State Lighting Multi-Year Program Plan (March 2010): http://apps1.eere.energy.gov/buildings/ publications/pdfs/ssl/ssl_mypp2010_web.pdf.
12DOE Solid-State Lighting CALiPER Program Summary Report Round 9 (2009).
13ENERGY STAR Program Requirements for Solid State Lighting Luminaires.
14Color Quality Scale http://colorqualityscale.com/.
15Color Rendering Index http://apps1.eere.energy.gov/buildings/publications/pdfs/ssl/ color_rendering_index.pdf.
16T. Komoda, H. Tsuji, N. Ito, T. Nishimori, and N. Ide, SID Symposium Digest 41, 993 (2010).
17M. Hack and P. A. Levermore, U.S. DOE R&D Workshop, Raleigh, NC (2010).
19P. Y. Ngai, U.S. DOE SSL Manufacturing Workshop, Fairfax, VA (2009). •