A Study on AlN thin film as Thermal Interface Material for high power LED

Transcription

1 International Journal of Electronics and Computer Science Engineering 296 Available Online at ISSN A Study on AlN thin film as Thermal Interface Material for high power LED Subramani Shanmugan *, Devarajan Mutharasu, Abu Hassan Haslan Nano Optoelectronics Research Laboratory, School of Physics, Universiti Sains Malaysia (USM),11800, Minden, Pulau Penang, Malaysia * Corresponding author shagan77in@yahoo.co.in (S.Shanmugan) Tel: ; Fax: Abstract- AlN thin film was coated over Al and glass substrates with 400 nm thickness using DC sputtering for thermal interface material (TIM) application. AlN coated Al substrate was used as a heat sink for 3W green LED. The thermal transient measurement was recorded for given LED attached with bare Al and AlN coated Al substrate at three different driving currents. The observed junction temperature (T J ) was low ( T J = 2.81 ºC) for AlN coated Al substrate at 700mA compared to bare Al. Thermal paste (TP) was also assisted to reduce the T J value but created the thermal barrier when applied on AlN thin film. Total thermal resistance (R th-tot ) was also noticeably decreased for AlN thin film. R th-tot values were high for AlN thin film coated Al substrates at lower driving currents (100 and 350mA) especially at TP condition. Interface resistance was low for AlN thin film as thermal interface material instead of TP. Keywords AlN, thermal interface material, Cumulative structure function, LED I. INTRODUCTION Solid-air interface represents the greatest barrier in thermal management. All surfaces have a certain roughness due to microscopic hills and valleys and poor surface flatness and 99% of the surfaces are separated by a layer of interstitial air between two typical electronic components [1]. Currently, the Light Emitting Diodes (LEDs) in the market have an efficiency of about 10% 20%. Consequently 80% 90% of the energy is converted into heat [2].The maximum light output, quality, reliability and the life time of LEDs are all closely related to the junction temperature. Thus, thermal management has become a key reliability issue for LED module and various solutions have been proposed [3]. The most commonly used LED board is Fire Retardant 4 (FR4). An alternative board with high thermal conductivities could be used for better thermal performance. The AlN has a theoretical thermal conductivity k TH of 320 Wm -1 K -1 [4], while the maximum experimental value for the bulk material is 270 Wm -1 K -1 [5]. However, the k TH of AlN film is much lower, in the range of Wm -1 K -1 [6, 7], since it depends on the deposition process details, grain size and shape, film thickness and impurity concentration. Nevertheless, it has been demonstrated that thermal conductivity and diffusivity of thin films can deviate significantly from its bulk form, due to variations in structural dimension and phonon transport mechanisms [8]. In particular, due to its electrical and thermal properties, AlN can be used for heat spreading/sinking in Integrated circuit applications. AlN is a material that is neither contaminating nor poisonous, differently from other materials, such as beryllia for example, and can be both deposited and etched by means already available in conventional silicon processing [9]. Very few works have been reported using AlN thin film as thermal interface material. But no attempts appear to have been made to study the cumulative structure function of 3W LED fixed on AlN thin film coated Al substrates as heat sinks as well as AlN as thermal interface material. In this study, thermal performance of 3W green LED on AlN thin film coated Al substrates as heat sink and the influence of TP on the T J and thermal resistance (R th ) is reported here. II. EXPERIMENTAL TECHNIQUE A. Synthesis and Thermal analysis of AlN thin film AlN thin films were deposited on Al (23cm x 25 cm) and commercial glass substrates (25 mm x 75 mm) using Al (99.99% purity) target (3 inch in diameter and 4 mm in thickness) by DC sputtering (Edwards make, Model-Auto 500). The chamber was initially evacuated to high vacuum 8.2 x 10-6 mbar by using a turbo molecular

2 A Study on AlN thin film as Thermal Interface Material for high power LED IV.297 pump backed by a rotary pump and fixed as base pressure for coating. High pure Ar (99.999%) and N 2 (99.999%) were used for AlN coatings. Ar and N 2 gas mixture ratio was fixed as 80:20 (16 sccm : 4 sccm with total of 20 sccm). The substrates were cleaned by rinsing in ultrasonic bath of acetone and isopropyl alcohol. All AlN thin films were coated at room temperature and the thickness of the film was 400 nm measured by digital thickness monitor. The deposition rate and sputtering power were kept constant at 0.6 Å / sec and 300 W, respectively. In order to remove the surface oxidation of the target, pre-sputtering was carried out for 5 min before starting deposition at Ar pressure of 3.2 x To get the uniform thickness, rotary drive system was used and 25 RPM was fixed for all AlN film coatings. All AlN thin films were coated at chamber pressure of 8.2 x Substrate to target distance of 7 cm was kept constant for all depositions. In order to test the performance of thin film interface material, AlN thin film coated Al and glass substrates were used as heat sink for 3W green LED package attached with Metal Core Printed Circuit Board. Cool master thermal paste kit was used as thermal paste to study the influence on thermal performance. The thermal transient characterization of the LED for different measurement conditions is captured (Table -1) based on the electrical test method JEDEC JESD-51. The thermal transient curve of the LED is captured by the Thermal Transient Tester (T3Ster) in still air box. B. K Factor Calibration Before the real measurement, the LED was thermally calibrated using dry thermostat and T3Ster as the power supply. The product of K and the difference in temperature-sensing voltage (referred to as V F ) produces the device junction temperature rise: T J = V F K (1) K = T J / V F (2) During the calibration process, the LED was driven with lower operating current at 1mA to prevent selfheating effect at the junction. The ambient temperature of the LED was fixed to 25 C and the voltage drop across the junction was recorded once the LED reaches thermal equilibrium with the temperature of the thermostat. Later, the ambient temperature of the LED was varied from 35 C, 45 C, 55 C, 65ºC, 75 C and 85 C and the voltage drop across the junction was noted at each ambient temperature. From the calibration process, the K-factor of the LED was determined (2.289) from the graph of junction voltage (voltage drop) against ambient temperature as shown in Fig.1. Figure 1. K factor calibration curve for given 3W green LED C. Thermal Transient Analysis During the thermal test, the LED was driven at three different currents 100 ma, 350mA and 700 ma in a still-air chamber at room temperature of 25 C ± 1 C. The LED was forward biased for 900s. Once it reaches steady state, the LED was switched off and the transient cooling curve of heat flow from the LED package was captured for another 900s. The obtained cooling profile of the LED placed over AlN thin film coated Al and plain Al substrates as heat sink was processed for structure functions using Trister Master Software.

3 IJECSE, Volume 2,Number 1 S. Shanmugan et al. IV.298 III. RESULTS AND DISCUSSION The thermal behavior of 3W green LED fixed over AlN coated substrates as heat sink for various driving current were measured using T3ster and the observed results are given in Table - 1. In order to check the influence of thermal paste, the thermal behavior was also tested in thermal paste condition. The observed cooling curves are given in fig. 2. Table 1 shows the results for Al substrates for various driving current. It reveals that the T J reduces noticeably for AlN coated substrates at 700 ma ( T J = 2.81 ºC). Instead of reducing the T J, the value slightly increases for AlN thin film coated substrates at lower driving currents (100 and 350mA). It is due to the molecular agitation (free electrons) within the material is small at low T J than the LED operates at high operating current [10]. Figure 2. Transient cooling curve of 3W green LED fixed on bare Al Figure 3. Cumulative structure function curve of 3W green LED and AlN thin film coated Al substrates with and without thermal paste fixed on bare Al and AlN thin film coated Al substrates with and measured at (a) 100 ma, (b) 350 ma and (c) 700 ma without thermal paste measured at (a) 100 ma, (b) 350 ma and (c) 700 ma Table 1. Thermal properties of 3W green LED on Al plate observed from cooling curve and cumulative structure function Al TP/Al AlN/Al TP/AlN/Al Driving Current (ma) T J (ºC) R th-tot (K/W) R th-b-hs(k/w)

4 A Study on AlN thin film as Thermal Interface Material for high power LED IV.299 In TP mode, the T J value drastically decreases for Al substrate ( T J = 3.84 ºC) when operated at 700 ma. The T J value is more for TP than AlN. It is attributed to the effect of micro contacts and it is more at TP condition when compared to solid AlN thin film. Moreover, the TP does not support on reducing the T J for AlN thin film coated Al substrates even operated at 700 ma. It is because of the composition of too many interfaces between MCPCB and Al substrate [11]. But TP does not show much effect on T J for lower driving currents (100 and 350mA). It is due to the formation of thermal barrier from the TP between two high thermal conductivity material contacts. From the transient curve, the cumulative structure function was also recorded as seen in fig. 3 (a-c) and the total thermal resistance (R th-tot ) was also measured form the cumulative structure function as given in Table -1. As reported in earlier paragraph, the R th-tot value slightly increases when operated at lower driving currents (200 and 350mA). It is may be due to stress or strain developed during the synthesis of AlN thin film over Al substrates and hence the lattice mismatch is possible [15]. It is considered as defected crystals grown at room temperature and influences the thermal conductance. From the fig. 3, the red color curve (AlN on Al substrates) shows high resistance as well as low capacitance value (see the change the thermal path at near 15 K/W). It reflects the interface resistance for these samples. In addition, a small and broad peak is also observed in the cumulative structure function ( see in between 20 and 25 K/W) for all samples and the curves belongs to AlN/Al substrates (Red and green color curves) deviates towards downwards from the curves of Al substrates at around R th of 20 K/W (black and blue color curves). It is only due to the mismatch at the interface between AlN and Al substrates. It is also evidently proved by observing the deviation in thermal path for 3W LED tested over AlN thin film coated glass substrates as heat sink (see fig.4b). Noticeable decrease in R th-tot could be observed for AlN coated Al substrates at 700mA. As discussed earlier, the TP supports to reduce the R th value for the LED fixed at Al substrates at higher driving current (700mA). Since the heat transfer through any material depends on the temperature difference between the hot and cold junction, the material transfers the heat at higher rate for high T. The same rule will apply for this LED package and high heat transfer is possible at high T J value and hence higher operating current shows low R th value with respect to T J. The high R th-tot value for AlN coated Al substrates at 350 and 700 ma when compared to TP is may be due to the reduced micro contacts between solid MCPCB and AlN thin film. But TP behaves as a liquid and fills all gaps between the contacting asperities and hence the resistance decreases. It is also observed that the increased R th-tot for TP used as an interface in between MCPCB and AlN coated Al substrates. It is attributed to the thermal conductivity mismatch between the TP ( Wm -1 K -1 ) [11] and AlN ( Wm -1 K -1 ) [6]. Because of this mismatch, thermal barrier occurs and restricts the heat flow from MCPCB to AlN thin film. Using the cumulative structure function, the interface thermal resistance (R th-b-hs ) was also measured and given in table 1. It shows that the value represents the heat transfer behavior of interface material (AlN or TP) between MCPCB and heat sink. From the table -1, it clearly indicates that the AlN interface material shows low R th-b-hs value ( R th-b-hs = 1.88 K/W) than the value for TP ( R th-b-hs = 1.45 K/W) as interface material at all driving currents especially at 700 ma when compared with Al substrates as heat sink. It may be due to the effect of temperature on the thermal conductivity of AlN thin film [13]. As grown Al thin film may have disordered structure at the interface region and hence the lower conductivity occurs. It is nullified by heat the material as a result of high junction temperature and the thermal resistance decreases at high operating current (700mA). But the R th-b-hs value shows high value for AlN coated Al substrate when TP applied at interface between MCPCB and AlN/Al especially at 100 ma (29.09 K/W). In addition, the behavior of AlN thin film as interface material is also tested and the cumulative structure function for 3W LED on AlN coated glass substrates as heat sink is measured at two different driving currents (100 and 200mA) is given in fig.4 (a&b). Fig. 4a shows the cooling curve of 3W green LED for the two different driving currents. It reveals that the T J value reduces noticeably for AlN thin film as interface material. The measured T J was 0.64 ºC and 1.45 ºC for 100 ma and 200 ma respectively. This study also insists the behavior of AlN thin film as interface material for thermal management in high power LEDs.

5 IJECSE, Volume 2,Number 1 S. Shanmugan et al. IV.300 Figure. 4. (a) Transient cooling curve and (b) Cumulative structure function curve of 3W green LED fixed on bare and AlN coated glass substrates measured at 100 and 200 ma. IV. CONCLUSION AlN thin film was used as thermal interface material for high power LEDs and the thermal performance was tested using transient curve analysis. The observed T J was low for AlN coated Al substrates at high driving currents (700mA). Interface thermal resistance in between MCPCB and AlN was also influenced the heat flow at lower driving currents (100 and 350 ma). TP was not supported to enhance the thermal conductivity on AlN/Al. Thermal resistance of the AlN interface materials was low at all currents when compared to TP. V. REFERENCE [1]. de Sorgo M, Thermal Interface Materials, Electronics Cooling, September, 122 (2), 15 (1996) [2]. Petroski J. Thermal challenges facing new generation light emitting diodes (LEDs) for lighting applications. SPIE Proceeding, 2002: 215 [3]. Huaiyu Y, Koh S, van Zeijl H, Gielen A W J, and Guoqi Z, A review of passive thermal management of LED module J. Semicon.,, vol. 32, pp. 1-4(14008), [4]. Slack G A, Nonmetallic crystals with high thermal conductivity, J. Phy..Chem. Sol., vol. 34, pp , [5]. Watari K and Shinde S L, High thermal conductivity materials, MRS Bulletin, vol. 26, pp , [6]. Kuo P K, Auner G W, Wu Z L, Microstructure and thermal conductivity of epitaxial AlN thin films, Thin Solid Films, vol. 253, pp , [7]. Lee J W, Cuomo J J, Cho Y S, Keusseyan R L, Aluminum nitride thin films on an LTCC substrate, J. Amer. Ceram. Soci., vol. 88(7), pp , [8]. Brotzen F R, Loos P J, Brady D P, Thermal conductivity of thin SiO2 films, Thin Solid Films, vol. 207, pp , [9]. [10]. [11]. [12]. Shanmugan S. Mutharasu D, Anithambigai P, Teeba N and Abdul Razak I, Structural properties of DC sputtered AlN thin films on different substrates, communicated. [13]. Pan T S, Zhang Y, Huang J, Zeng B, Hong D H, Wang S L, Zeng H Z, Gao M, Huang W, and Lin Y, Enhanced thermal conductivity of polycrystalline aluminum nitride thin films by optimizing the interface structure, J. Appl. Phys., vol. 112, , 2012