EFFECT OF DIFFERENT MICRO STRUCTURES ON HEAT DISSIPATION OF CPU SURFACE HEAT SINK

Transcription

1 SCIENTIFIC PAPERS EFFECT OF DIFFERENT MICRO STRUCTURES ON HEAT DISSIPATION OF CPU SURFACE HEAT SINK Yan WANG 1,a ABSTRACT: In order to discuss the heat dissipation perormance o CPU heat sink, based on the commonly used CPU Aluminum Alloy heat sink,experiment simulation device o CPU heat sink and wind tunnel experiment platorm are established. On the basis o comprehensively considering convection heat dissipation and radiation cooling, the eect o surace micro structure and radiation coating on the CPU thermal perormance is studied. The results show that ater the surace treatment and coating, the degree o rough surace increases, and the surace temperature o heat sink reduces about 5 DEG C. By using the composite treatment by oxidation ater spray coating, the cooling eect is signiicant. Compared with the direct spray, when the input power is 9W, the cooling rate is nearly 1 DEG C, which can be a good solution to the CPU heat problem. As a result, it is concluded that it is o great signiicance in improving the electronic equipment reliability and saety perormance. KEY WORDS: CPU heat sink; surace processing; roughness; radiation rate. 1 INTRODUCTION With the rapid development o integrated circuit technology, all kinds o electronic equipment have been developed to high requency and high integration. The volume o electronic devices and products is getting smaller and smaller, and the heat lux around the integrated devices is becoming larger and larger. With the computer CPU as an example, the chip volume decreases constantly, and its integration rapidly increases. At last, it results in that in the running process, CPU chip heat lux increased sharply, while the electronic device is highly sensitive to temperature. I the chip heat has not dissipated in time, it will aect the lie and the working reliability o electronic devices [1]. Thereore, the eective solution to the problem o heat dissipation can promote the extensive application o high-density integrated technology and accelerate the integration o micro electronics industry. 2 LITERATURE REVIEW The common way to dissipate heat is to add heat sinks to the CPU, and how to optimize the heat sink to dissipate the heat o the CPU through the heat sinks so as to ensure the CPU work at normal temperature has become the ocus o the scholars. Sohel and his partners [2], through the experiment, proved that applying Al 2 O 3 nano luid spraying the surace o the heat sink can signiicantly improve 1 Department o Inormation Techonology, Changchun Finance College, Changchun, China a wdvw32@163.com. the perormance o heat transer. Moraveji [3] made use o CFD luid dynamics modeling method to characterize the thermal perormance o Al 2 O 3 nano coating rom a single and whole stage. In addition, Bahiraei proposed a new bio nano luid, so as to improve the cooling rate o the heat sink ins, to ensure the normal operation o equipment []. Based on this, this paper uses chemical grinding, chemical oxidation, chemical polishing, shot peening, surace spraying and other processing methods or the aluminum heat sink surace treatment. And then the micro structure with dierent roughness can be obtained. The surace radiation rate and heat measurement eect are measured ater the surace treatment, and actors aecting the cooling perormance o the radiator are analyzed, so as to improve the reliability and saety o operation o electronic equipment. 3 METHOD 3.1 Heat dissipation mechanism o heat sink Heat dissipation way o heat sink is usually orced convection and heat radiation. The orced convection uses an driven orce to orm the wind ield to strengthen air convection, which is the main way or heat dissipation. The heat radiation is that the heat dissipates to the surrounding environment rom the heat sink by way o inrared radiation. The main way o heat transer includes heat conduction, heat convection and heat radiation three kinds. In this study, CPU transer the heat produced to the heat sink in the heat conduction method. The aluminum heat sink transers the heat to the space ACADEMIC JOURNAL OF MANUFACTURING ENGINEERING, VOL.15, ISSUE 3/217 29

2 SCIENTIFIC PAPERS mainly by way o thermal convection and thermal radiation. The ormula o heat exchange is considered when convection heat [5] is considered separately: Q d At A t c c ( w t In the above ormula, tw suggests the object surace temperature ( ); t indicates the temperature ( ) o luid; Δt reers to the temperature dierence ( ) between the object surace and the luid; A represent the surace area (m2) o the object involved in heat transer (); and means convective heat transer coeicient (W/m2 ). When the volume o heat sink is determined, i the contact volume with the luid is increased, the heat exchange can be increased. When the radiation heat radiation is considered separately, the heat emitted by the radiating surace o the radiating in can be expressed by the ollowing ormula [6]: Q r A( t w t In the above ormula, ε reers to radiation rate o an object surace, whose value is related to material temperature, surace area and surace roughness; δ is Stepan - Boltzmann constant, whose value is W/(m-2 K-). The heat sink ins can be approximately considered as a rectangular structure. The heat dissipation o the heat sink is deducted by using the ininitesimal method. The heat conduction along the Solution composition Na 2 CO 3 ~ Na 2 CrO 1~2 NaOH 2~3 ) ) Content (g/l) rib elongation direction, the temperature o rib is t. The temperature o heat sink at a certain place is t, the thermal conductivity coeicient is, and the perimeter o cross section is U. The quantity o heat can be calculated by the ollowing ormula: Q A U( t c t ) 2AU ( t t t t1 ) According to this ormula, it is known that, the heat dissipation o the radiating in is related to the temperature dierence, the surace area o the radiating in, the radiation rate and the coeicient o thermal conductivity. 3.2 Surace processing In general, in the CPU radiator, heat sink is made by aluminum alloy by extrusion technology. The aluminum radiator is made by die casting process, and the surace treatment must be done uniied with the surace pretreatment. The heat sink with the same size o 19 copies is made, and divided into 6 groups. The irst ive groups have three blocks, used or micro structure processing. And the last group has our blocks, applied or radiation coating thickness test. The irst step o the surace pretreatment is to make use o alkaline cleaning agent (compositions are shown in table 1) or oil removal. Followed by, nitric acid or phosphoric acid is used or pickling, and inally, the samples are washed and dried in the lowing water, which can orm the clean and uniorm surace. Response time (min) Table 1 Surace cleaning ormula and process Temperature ( ) 3~5 ~ 1 2 Chemical oxidation treatment [7]: weigh 5g Na 2 CO 3 and 15g Na 2 CrO to ill in the beaker with boiling water, place the sink into the 9~ DEG C solution, and continuously shake or 5~1min. Ater a layer o gray ilm is ormed on the surace, the specimen is removed. In the weak acid solution, the surace residual oxide is cleaned, and then it is lushed in the cold water, or drying in the air. Chemical grinding processing: weigh 3~32g/L NaOH, ~5g/L Na 3 PO, and 15~25g/L NaF, respectively. Add into the boiling water, and the control solution temperature is ~7 DEG C. When the sample surace appears obvious sand surace eect, the specimen is removed or cleaning in the weak acid. Ater it is cleaned, the sample is rinsed with cold water until the surace is cleaned, or drying in the air. Chemical polishing treatment: place the aluminum alloy material in the polishing liquid with reasonable ratio o corrosion agent, oxidizing agent, and additives. The time is controlled in 1~12s, and the sample surace has a series o chemical reactions, which induce the morphological changes. Mechanical shot peening treatment: use stainless steel bullet in compressed air to blow the surace o the radiator, to make the material surace 3 ACADEMIC JOURNAL OF MANUFACTURING ENGINEERING, VOL.15, ISSUE 3/217

3 SCIENTIFIC PAPERS appear cyclic plastic deormation. Generally, it requires sample surace coverage rate more than %. Spraying treatment: by using the manual cold spraying technology, the double volatile coating is coated on the surace o the heat sink, orming a layer o bright, clean and uniorm coating. In order to explore the eect o dierent thickness o the coating o radiation coatings on the cooling eect, in the experiment,.1,.15,.2, and.25 our kinds o coating thickness are chosen. 3.3 Measurement o surace roughness o samples The surace roughness reers to the micro geometrical characteristics composed o the smaller spacing and peaks and valleys on the machined surace. The smaller the surace roughness is, the smoother the surace o the sample is [8]. The roughness measurement uses the middle line system, which takes the contour middle line as the measuring datum line. The sampling length L is used or distinguishing a base line length with the surace roughness characteristics, taking the horizontal direction. The assessment length o Ln is applied or assessing a section length required or the contour assessment. In general, within an assessment length, 5 sampling lengths are chosen. The surace roughness o the central system commonly uses three ways, including the arithmetic average deviation o the contour Ra, the micro roughness height Rz and the maximum height o the contour Ry. In this experiment, the surace roughness o the sample is measured by Time Inc TR2 hand-held roughness tester. The working low chart is shown in Figure 1. The nine points with uniorm distribution are selected on the sample surace, so as to reduce the measurement error. Electromagnetic absorber Mechanical transmission equipment Sensor Motor Relay Power supply 3. Heat sink perormance test The radiation rate is a measurement o the ability o object surace to release the energy in the orm o radiation. The object radiation ratio is equal to the ratio o black body radiation energy and energy o the object radiation at a certain temperature at the same temperature. The radiation rate o actual object is related to the surace state o the object, including surace temperature, surace roughness, surace oxide layer and so on [9]. In order to measure the inluence o surace treatment on the perormance o CPU heat sink, a wind tunnel test platorm was built to test the surace temperature o the radiator under the orced convection. In the experiment, the heating rod is used as the heat source to simulate the CPU chip to provide heat or the heat sink. A simple wind tunnel is made o a high-power an and an aluminum alloy plate, and the wind speed at the exit o the wind tunnel is about 15m/s. The heat radiation simulation device is arranged in a wind tunnel, the distance to the an is about 3mm, and the airlow is evenly distributed through the heat sink to take away heat. When the input power is changed, the heat sink is used to collect the surace temperature o the heat sink. Each point is measured several times and the average is at last taken. RESULTS AND DISCUSSION.1 Test results o sample surace radiation rate In the experiment, the radiation rate o the sample ater surace treatment is measured. The surace temperature o each sample is about degrees centigrade, the ambient temperature is 26.5, and the humidity is 5%. The surace radiation rate o each sample is shown in igure 2. Ater the surace treatment, the surace roughness o the sample increases, and the radiation rate increases with the increase o the surace roughness. Compared with the original surace, the radiation rate increases 1.5 times ater oxidation and the change is obvious. Trigger Pre-level Ampliier A/D conversion Wavelength and length o assessment DSP chip Figure 1 Flow chart o roughness instrument ACADEMIC JOURNAL OF MANUFACTURING ENGINEERING, VOL.15, ISSUE 3/217 31

4 Temperature( ) Emissivity Temperature( ) SCIENTIFIC PAPERS Heat sink A Heat sink B Original sample Chemical polishing Chemical scrub Mechanical shot peening Chemical oxidation Sample Figure 2 Surace radiation rate o each sample The coating thickness is about 7μm in the experiment. The coverage rate reached %, the surace brightness sample increases obviously ater coating, and the surace radiation rate calibrated at DEG C is about.98, close to the black body. The surace spraying signiicantly increased the sample radiation rate. Ater surace treatment, the surace o specimen is also coated with a radiation dissipation coating, and the radiation rate is measured. The results show that the radiation rate o the surace o the sample is close to 1. That is to say, when the coating thickness is large enough that the coverage rate is over %, the substrate surace morphology o the sample has little eect on radiation rate Figure 3 The temperature data beore the correction The sotware origin is used or itting o the experimental data ater being corrected, and the results are shown in igure. The regression equation distribution is shown as ollows: Y Y A B X X The absolute coeicient R 2 is close to 1, and the itting degree is optimal. The curve showed that: ater spraying the heat radiation coating, the surace temperature o the heat sink is reduced by about 5 DEG C. The slope o two regression models is basically equal, and the surace cooling eect is basically stable with the increase o input power. A B.2 Inluence and its correction o environmental temperature on the experimental results When the heat sink enters the stable temperature, as the temperature increases by 2 degrees centigrade or the same sample, the data measured twice is about ~5 DEG C, which seriously aects the accuracy o the experiment, so it is necessary to correct the experimental data. With a certain linear relationship with environmental temperature and heat sink temperature at the bottom, so the linear itting can be used to the linear equation o heat sink temperature changing with the ambient temperature. The temperature o all test points is corrected as that in the benchmark ambient temperature o 25 DEG C, and then comparative analysis is made. A represents the spray not coated, and B reers to the heat dissipation coating with the surace heat coating thickness about 7μm. The experimental data beore the amendment is shown in igure ACADEMIC JOURNAL OF MANUFACTURING ENGINEERING, VOL.15, ISSUE 3/217 Heat sink A Heat sink B Figure The experimental data o heat sink ater the correction.3 Eect o coating thickness on heat dissipation result In the experiment, our kinds o coatings with dierent thickness are sprayed on the surace o the heat sink. The coating thickness is control in.1mm,.15mm,.2mm, and.25mm, respectively. In the experimental device, the highest ground temperature o heat sink is measured. The radiating simulation device is arranged on the wind tunnel. Ater repeated measurements, the measured data are corrected to the reerence ambient temperature o

5 Ra(μm) Heat sink surace temperature( ) Heat sink surace temperature( ) SCIENTIFIC PAPERS 25 DEG C. The measured temperature value is shown in igure 5, and the linear itting conidence is set to 95%. From the itting curve, it can be seen that, the increase o the coating thickness limits the cooling o the heat sink, increases the thermal resistance, and increases the temperature mm.15mm.2mm.25mm 2 Figure 5 The surace temperature o the heat sink varies with the thickness o the coating ater the correction. Inluence o surace roughness on temperature o heat sink Dierent micro morphologies were processed on the surace o the sample and their roughness was measured. The results are shown in igure 6, and it can be seen that, the roughness was increased obviously ater surace treatment Original surace Polished surace Chemical scrub Shot peening Chemical oxidation Sample Figure 6 Ra value o surace o heat sink ater dierent surace treatments Any object is able to make inrared radiation to the space. The physical parameter or the measure o the object size is the radiation rate. That is the ratio o electromagnetic energy radiated outwards through the surace to the electromagnetic energy o black body under the same conditions. The radiation rate is related to the material structure, physical and chemical properties, the temperature and so on, to a great extent, it also depends on the surace state o the object. The CPU chip transers heat to the ins, which dissipate heat into the space in the orm o orced convection and inrared radiation. For the inrared radiation, the sample ater surace treatment, the surace roughness increases, which showed irregular topography with uneven surace, namely the surace roughness. Such kind o irregular surace structure has a great eect on the radiation rate. Dierent surace roughness sample temperatures, ater repeated measurements and corrections, the average value is shown in igure 7. Ater the surace treatment, the heat sink temperature decreased slightly, and the biggest drop is about 5 DEG C. With the increase o surace roughness, the heat sink temperature decreases slightly. In the igure, the slope o the regression models have no great change. I the input power continues to increase, the temperature change amplitude will not be too big Original surace Polishing Scrub Shot peening Oxidation 2 Figure 7 impact o surace roughness on the heat sink temperature.5 Inluence o spraying treatment ater substrate roughening on the perormance o heat sink The heat sink ater surace treatment, the radiation heat dissipation coating with the thickness about 7μm is coated on the surace. The surace temperature in the wind tunnel is tested. Ater correction, the data is shown in igure 8. The linear regression is made in origin, and the conidence is 95%. The absolute coeicients R 2 o each curve are all close to 1, and the itting degree is optimal. As can be seen rom the diagram, the surace treatment has certain inluence on the heat radiation coating, and the higher the input power is, the more obvious the eect is. When the input power is 9W, the temperature can be reduced by nearly 1 degrees at most. Ater treatment, the surace radiation rate o the sample does not increase, but the surace roughness increases the surace area between the radiator surace and the heat radiation coating, which is beneicial or the heat transerring rom the heat sink to the coating. ACADEMIC JOURNAL OF MANUFACTURING ENGINEERING, VOL.15, ISSUE 3/217 33

6 Heat sink surace temperature( ) SCIENTIFIC PAPERS Original surace Polishing Scrub Shot peening Oxidation intererence actors. In addition, a detailed theoretical analysis o the actors inluencing the radiation rate and coating perormance is made, so as to play an important guiding role in the optimization design and actual application o radiator Figure 8 The heat sink surace temperature itting curve ater the correction 5 CONCLUSION This paper, in terms o aluminum alloy CPU heat sink, uses the way o combining theoretical analysis and experimental study, to explore the eects o surace treatment technology on heat sink perormance. With chemical polishing, chemical mechanical grinding, chemical oxidation, and shot peening, the surace o the samples is processed. Changing the micro structure o specimen surace, the surace roughness is measured. The original sample is.27μm, chemical polishing is 1.2μm, chemical scrub is.6μm, mechanical shot peening is 5.63μm, and the surace roughness is 16.5μm ater chemical oxidation. The regression analysis is used or correcting the inluence o environmental temperature. All temperature data are corrected to the reerence ambient temperature o 25 DEG C or analysis and comparison. The greater the surace roughness is, the greater the emission rate is, and the lower the surace temperature is. Compared with the original sample, ater chemical oxidation, the surace temperature o heat sink is reduced by about 5 DEG C. The spraying surace radiation coating can eectively increase the radiation rate o the heat sink surace, so the surace temperature is reduced about 5 DEG C. The surace roughness will strengthen the radiation dissipation cooling perormance o coatings. The greater the surace roughness is, the higher the temperature is, and the more obvious the strengthening eect is. Ater oxidation, the most obvious the cooling eect o the spraying coating is, and the cooling rate is nearly 1 degrees. In this paper, the characterization o sample surace structure ater the surace treatment is limited to the surace roughness. The urther comprehensive analysis o the change o the sample is not carried out. And the external environment will greatly aect the measurement precision and structure. It is necessary to urther design high precision measurement platorm, or eliminating 6 REFERENCES Liu, Di, et al. Thermoelectric mini cooler coupled with micro thermosiphon or CPU cooling system. Energy 83 (215): Sohel, M. R., et al. An experimental investigation o heat transer enhancement o a minichannel heat sink using Al 2 O 3 H 2 O nanoluid.international Journal o Heat and Mass Transer 7 (21): Moraveji, Mostaa Keshavarz, and Reza Mohammadi Ardehali. CFD modeling (comparing single and two-phase approaches) on thermal perormance o Al 2 O 3/water nanoluid in minichannel heat sink.international Communications in Heat and Mass Transer (213): Bahiraei, Mehdi, and Saeed Heshmatian. Application o a novel biological nanoluid in a liquid block heat sink or cooling o an electronic processor: Thermal perormance and irreversibility considerations. Energy Conversion and Management 19 (217): Zhou, Mingdong, et al. Industrial application o topology optimization or combined conductive and convective heat transer problems. Structural and Multidisciplinary Optimization 5. (216): Albella, Pablo, et al. Electric and magnetic ield enhancement with ultralow heat radiation dielectric nanoantennas: considerations or surace-enhanced spectroscopies. Acs Photonics 1.6 (21): Nieh, Hwa-Ming, Tun-Ping Teng, and Chao- Chieh Yu. Enhanced heat dissipation o a radiator using oxide nano-coolant. International Journal o Thermal Sciences 77 (21): Kumar, Ravinder, and Santram Chauhan. Study on surace roughness measurement or turning o Al 775/1/SiCp and Al 775 hybrid composites by using response surace methodology (RSM) and artiicial neural networking (ANN). Measurement 65 (215): Branch, Ruth, C. Chris Chickadel, and Andrew T. Jessup. Inrared radiation rate o seawater and oam at large incidence angles in the 3 1μm wavelength range. Remote Sensing o Environment 18 (216): ACADEMIC JOURNAL OF MANUFACTURING ENGINEERING, VOL.15, ISSUE 3/217