Investigation of the long-term aging characteristics of chip-on-board LEDs

Chip-on-board (COB) light-emitting diodes (LEDs) are a relatively new LED package platform
that provides the highest density lighting flux with the thinnest profile. COB LEDs are formed by
interconnecting large numbers of mid-power LED (MP-LED) die to form a dense light-emitting
surface (LES). Because the number of LEDs used in the COB LED packaging platform is very
large (usually between 12 and 100), more light and waste heat per unit area are produced relative
to other LED package platforms. Therefore, it is necessary that the MP-LED die are situated
directly on an excellent thermally conductive substrate. The COB-LED assembly is covered by a
large silicone–phosphor layer creating a phosphor-converted LED (pc-LED).
This report, which is the second in a series of reports about the reliability of commercial COB
LED products, focuses specifically on the long-term aging characteristics of COB LEDs. All five
COB-X products examined in this study are from different Tier 1 manufacturers and represent a
myriad of LES sizes, die interconnection methods, substrates, and construction techniques. All
COB products in the test matrix are pc-LEDs, and the LES of the devices under test (DUTs)
varied from 4.5 millimeters (mm) to 14.6 mm. Most DUTs were set to a fixed correlated color
temperature (CCT), but one COB (i.e., COB-3) had the ability of white tuning. Three of the COB
products examined during this study were built on metal-core printed circuit boards (MC-PCBs),
and the remaining two COB products were built on ceramic substrates. Four COB products
contained LED die that were interconnected by wire bonding, and the LED die in these products
were mounted to the substrate with a thermally conductive adhesive. The LED die in one product
was flip chip–bonded to an interposer layer by using solder bumps, and the interposer board was
mounted on an aluminum nitride (AlN) substrate for thermal management purposes. The number
of LEDs in the LES ranged from 12 (for a 4.5-mm LES) to 108 (for a 14.6-mm LED), and four
of the five products had a blue LED pump, and one product had a violet LED pump. The radiant
efficiency of these devices ranged from 0.31 to 0.44, and the luminous efficacy of these samples
ranged from 83 lumens per watt (lm/W) to 127 lm/W.
This report builds upon the aging data from the previous report [1]. Prior to testing the COB
DUTs, the samples were mounted on appropriate heat sinks intended to keep the case
temperature (Tc) of the COB package within the manufacturer’s specifications. Three different
accelerated stress tests (ASTs) were used: room-temperature operating life (RTOL), an elevated
ambient environment of 75 degrees Celsius (°C) and 75% relative humidity (7575), and
operation in an oxygen-free (anaerobic) environment. The test population for RTOL was
expanded from 3 DUTs from the previous report [1] to 10 DUTs for each COB product to
comply with American National Standards Institute (ANSI)/Illuminating Engineering Society
(IES) technical memorandum (TM) TM-21-19. The test population for 7575 was also set to 10
DUTs, whereas the test population for the anaerobic environment was kept to 2 DUTs because of
space limitations in the anaerobic chamber. The RTOL and 7575 DUTs completed 6,000 hours
(hrs) of testing, whereas the anaerobic DUTs completed 2,000 hrs.
In the context of this report, reliability is judged by either abrupt (i.e., lights-out) failures caused
by electrical open circuits or parametric failures caused by excessive luminous flux depreciation
Investigation of the Long-Term Aging Characteristics of Chip-On-Board LEDs: Operating Lifetime Study
viii
or chromaticity shifts. The luminous flux maintenance (LFM) data were analyzed by using an
exponential decay model as indicated in ANSI/IES TM-21-19. The time to reach an LFM of 0.7
(i.e., L70) was projected by using the exponential model in ANSI/IES TM-21-19. During RTOL
testing, the projected time to L70 for all COB products (except for Product COB-5, in which L70
= 32,200 hrs) was limited by the 5.5 times rule (33,000 hrs for 6,000 hrs experimental time and
10 DUTs), and the largest magnitude of chromaticity shift (Δu’v’) was 0.0036 for Product COB5. These results indicate good reliability of the COB products in the RTOL environment. No
abrupt or parametric failures were observed through 6,000 hrs of the RTOL tests.
The 7575 tests proved to be a higher stress environment that produced TM-21-19 decay rate
constant (α) values significantly higher than those measured during RTOL tests. The α values for
the 7575 test were also significantly higher than the α values found in the previous report [1] for
testing conducted in an elevated ambient environment of 75°C (75OL). The Tc values of the
7575 test were slightly higher than the RTOL test but not significantly different than the 75OL
test, so it is most likely that the additional moisture in the 7575 test environment was responsible
for the larger α values. The accelerating nature of the 7575 test led to 35 parametric failures (out
of 53 total DUTs) during 6,000 hrs of 7575 exposure because of excessive chromaticity shift
(i.e., Δu'v' ≤ 0.007) and 20 parametric failures that also exhibited excessive luminous flux
depreciation (i.e., LFM < 0.7). All DUTs that failed parametrically because of low LFM also
failed because of excessive chromaticity shift. The parametric failures occurred only in some
products: all DUTs for Products COB-2, COB-3, and COB-5 failed parametrically in one or
more areas, but Product COB-6 did not experience any failures. Product COB-4 was removed
from testing after 4,000 hrs because of cracking of the silicone LES and discoloration of
electrical leads, which were soldered on prior to purchase. During testing, it appeared that the
leads were unable to handle the forward current load of the COB.
Because of the high-photon-induced darkening effects observed by the lighting industry for some
high-power LEDs (HP-LEDs) in low-oxygen environments and the high flux nature of COBs,
this study examined whether darkening effects are also observed for COB DUTs in an anaerobic
environment. Product HP-1, an HP-LED known by the industry to undergo LES darkening in the
anaerobic environment, was used to validate the anaerobic tests. For Products COB-4 and COB5, LFM was greatly reduced by 240 hrs of operation with LFM less than 0.80 being measured;
however, the LFM reduction was not as significant as it was for Product HP-1 (LFM = 0.18 at
240 hrs). By 2,000 hrs, the LFM for 6 of the 10 COB DUTs in the anerobic chamber fell below
the parametric failure threshold; the DUTs also exhibited excessive chromaticity shift. For
Product COB-5, the anaerobic environment was approximately 16 times more accelerating than
the 7575 environment even though the DUT’s junction temperature (Tj) value was lower because
of the use of the cooling plate in the anerobic environment. Overall, the chromaticity shift and
LFM degradation were found to be reversible for the COB DUTs (and Product HP-1) after 100
hrs of operation in air.
A high level of correlation (i.e., absolute value of the correlation coefficient r > 0.8) was found
between current density, flux density, LFM, and Δu'v' for DUTs operated in the r anaerobic
Investigation of the Long-Term Aging Characteristics of Chip-On-Board LEDs: Operating Lifetime Study
ix
environment. For these samples, LFM was found to be negatively correlated to current density,
flux density, and Δu'v', whereas current and flux densities were found to be positively correlated
to Δu'v'. In contrast, for RTOL, a positive correlation was found between LFM and Δu'v' and
current density. This correlation is believed to be the result of higher external quantum efficiency
(EQE) of the LED (ηLED,EQE) for products capable of operating at higher current densities. Weak
or no correlations were observed for the DUTs operated in the 7575 environment, indicating that
the high stress environment of added temperature and humidity might outweigh effects because
of increased current or flux densities.
All blue pump LED DUTs examined during this study exhibited either shifts in the blue or green
direction or a combination of the two during the test period; either shift is indicative of
chromaticity shift mode (CSM) behavior that can be classified as CSM-1 or CSM-2. CSM-1 is
usually caused by a relative increase in emissions from the LED emitter, whereas CSM-2 is
usually caused by a change in the emission spectrum of warm white phosphors. The analysis
findings presented in this report indicate that the chromaticity shift of the COB LED products
examined during this study was primarily because of package effects arising from the silicone–
phosphor composite in the LES. These effects can include increased absorption of light rays with
long optical path lengths or chemical changes in the phosphor particle due to the high moisture
permeability of silicones. Although contributions from other package factors such as the
oxidation of silver (Ag) mirrors cannot be fully eliminated, especially for Products COB-2,
COB-3, and COB-5, it is more likely that changes in the silicone–phosphor LES was the primary
driver in chromaticity shifts and luminous flux loss.
COB LEDs are an emerging LED packaging platform for solid-state lighting (SSL) products that
offer products in high lighting density and in a thin profile. During RTOL testing, the LFM
performance of the COB LED products was excellent, with most project lifetime values limited
by the 5.5 times rule of ANSI/IES TM-21-19. For 7575, however, 66% of the DUTs failed
parametrically because of excessive chromaticity shifts by 6,000 hrs. These findings are
consistent with the highly accelerating nature of the 7575 testing environment and demonstrate
its utility in reducing the experimental time required for failure so that the failure modes can be
studied. Furthermore, for the anaerobic test, 60% of the DUTs failed parametrically by 2,000 hrs.
The high light density and thin profile of the COBs may help to create new SSL products with
advanced lighting features such as tunable white or unique optical patterns for light delivery.
However, COB LED packages also present challenges because of silicone expansion and
swelling due to moisture ingress and darkening effects because of high flux. Using AST methods
to understand these failure modes and developing new materials and manufacturing methods to
overcome these limitations will improve both the energy efficiency and reliability of the COB
LED packages in demanding high light–intensity applications. Ultimately, these types of studies
help promote energy efficient technology and reduce carbon emissions from lighting systems