- Email: [email protected]
Performance analysis and assessment of thermoelectric micro cooler for electronic devices
Energy Conversion and Management 124 (2016) 203–211
Contents lists available at ScienceDirect
Energy Conversion and Management journal homepage: www.elsevier.com/locate/enconman
Performance analysis and assessment of thermoelectric micro cooler for electronic devices Yang Cai a, Di Liu a,⇑, Fu-Yun Zhao b, Jian-Feng Tang a a b
College of Pipeline and Civil Engineering, China University of Petroleum, 266580 Qingdao, Shandong Province, China School of Power and Mechanical Engineering, Wuhan University, Wuhan, Hubei Province, China
a r t i c l e
i n f o
Article history: Received 25 February 2016 Received in revised form 29 June 2016 Accepted 3 July 2016
Keywords: Thermoelectric cooling system COP e NTU Operating mode
a b s t r a c t A novel operating mode of thermoelectric module (TEM, cooling, heating, generation) is established for electronic devices cooling, based on the method of effectiveness-number of transfer units (e NTUÞ. This work mainly focused on the effect of thermoelectric properties and the scale of extender block on cooling performance under different operating conditions in order to obtain effective cooling operating mode. Based on the TEM parameters, two sets of analytical solutions for thermoelectric cooler (TEC) are derived for the chip temperature Tj at a fixed cooling power Qc and Qc at a fixed Tj, respectively. The performance of TEC with/without scale of extender block is studied for the lowest chip temperature and maximum cooling capacity at fixed conditions. Analysis results show the thermoelectric properties and extender block are significant characteristics for different operating conditions. However, the coefficient of performance (COP) and temperature difference changed a little under given thermoelectric properties. The results indicate TEC system applied in electronic devices obtains effectively cooling module by controlling operating parameters, which do not changed with scale of extender block. The validation of the present analysis is also conducted compared with previous studies and through the infrared thermal imager. Ó 2016 Elsevier Ltd. All rights reserved.
1. Introduction High power electronic devices with smaller size and higher encapsulation have resulted in the focus on the thermal management [1]. Investigations show the thermal reliability of a silicon chip decreases by about 10% for a rise of 2 °C in temperature [2], resulting in the increasing of thermal management challenges. Simultaneously, traditional passive cooling technology such as natural convection cooling, forced air cooling, liquid cooling, are reaching the limits of cooling capacity and cooling efficiency for high power electric devices [3–5]. Compared with traditional cooling, TEC is thought to be one of the preferable technologies due to its numerous advantages, such as small volume, environmentally friendly and temperature control capability [6]. Therefore, developing thermoelectric micro cooler system has potential to make important contributions to the growing thermal management challenges of electronic devices. Most of the previous work on the TEC system has examined optimization and performance improvement to electronic devices by building various thermoelectric modules [7–16]. Zhang [13] ⇑ Corresponding author. E-mail addresses: [email protected] (D. Liu), [email protected] (F.-Y. Zhao). http://dx.doi.org/10.1016/j.enconman.2016.07.011 0196-8904/Ó 2016 Elsevier Ltd. All rights reserved.
and Zhang et al. [14] presented a generalized theoretical model in evaluating and optimizing thermoelectric coolers. Their modules did not consider the heat exchange in heat sink through heat exchange medium. Thermal design method covering with e NTU for solving the heat dissipater problem of TEC system have been investigated [17–21]. It is noted that their module expressions were not involved the temperature-difference and operating mode of effective cooling module. At present, the operating mode of thermoelectric module in practical use have not developed because of the dynamic process covering with temperature difference with current. Only Hodes [7] investigated the operating mode of TEM under the operating parameters to illustrate the application of the analysis framework. However, for a fixed hot and cold side, the energy conversion is changing with the temperature difference, and the analysis is extended to practical use, such as electronic heat dissipater. To the best of author’ knowledge, most of previous studies are heat dissipater for electronic devices focused on thermoelectric configuration and optimizing parameters. The thermal design theory (e NTU) applied in thermoelectric cooling system have not been developed. This work focuses on the operating mode of TEM under the different scale of extender block for various thermoelectric properties. Based on the thermal design theory, the
204
Y. Cai et al. / Energy Conversion and Management 124 (2016) 203–211
Nomenclature A C COP I Ii In Isat Ish K Kex L Q Qc,open Qcmax R S T Tho Tj DT DT max U UA Z ZT
area (m2) specific heat (W K1) coefficient of performance current (A) current at open circuit (I = 0) value (Qc,open) (A) neutral current at DT ¼ 0 (A) current at Qc = 0 (A) short-circuit current (A) thermal conductance (W K1) thermal conductance of extender block (W K1) length (m) heat flux (W) cooling quantity with I = 0 (W) maximum cooling load (W) resistance (X) Seebeck coefficient (V K1) temperature (K) hot side reference temperature (K) surface temperature of chip (K) TEC hot side to cold side temperature difference (K) maximum temperature-difference(K) overall heat transfer coefficient (W m2 K1) heat performance parameter (W K1) figure of merit (1/K) dimensionless figure of merit
Greek symbols r heat loss efficient rhe;h heat loss efficient of heat exchanger rte heat loss efficient of thermoelectric cooler e heat exchanger efficient u equivalent thermal efficiency Subscript c ejc ex h he in jc jc,c jc,h max n sh te
TEC cold side equivalent cold side junction to TEC extender block TEC hot side hot-side heat exchanger inlet junction to TEC cold side junction to TEC hot side junction to TEC Maximum Neutral short-circuit thermoelectric cooler
expression of cooling power and chip temperature are obtained for considering various operating mode. Two sets of analytical solutions for thermoelectric cooler are derived for the chip temperature at a fixed cooling power and cooling power at a fixed chip temperature, respectively. The objective of operating mode of TEM is to establish effective cooling module to predict the efficiency of TEC system and set up references for TEC system. Comparison with previous work and the full-scale experimental are also conducted for verification of the present analysis method.
balance equation. Thus, the theoretical equation for the TEM can be given as follows [23,24],
2. Schematic diagram of TEC system
where I is the electric current, Th and Tc are the junction temperature of thermoelectric cooler. The Seekbeck coefficient S, thermal conductivity K and the total electrical resistance R are the physical properties of a TEM, which is important for the thermoelectric module. Thermoelectric cooler COP is defined as
As shown in Fig. 1, the schematic of TEC system consists of TEM, hot side heat exchanger, extender block, and heater block. The TEM comprises many pairs of P-N type semiconductor columns, metallic connectors, and two electrically insulting ceramic plates [22]. The interface face contacted with electrical device by extender block is the cold side of thermoelectric cooler. Inversely, the interface face connected with heat sink is the hot side of thermoelectric cooler, and in this paper the heat sink with fins is called as hot side heat exchanger. The extender block, which could avert decreasing the cooling capacity and COP, has been considered as a buffer plate. A hot side heat exchanger with fin is used to take away the heat from heat source in the hot-side of TEC. The design task is to minimize the chip temperature or maximum cooling thermoelectric cooling load through TEC enhanced air cooling or others cooling ways. It is noted that the hot or cold side of TEC is not dependent on the temperature level. The contact of thermoelectric two side and aluminum extender is coated with thermal grease for decreasing thermal resistance. An axial fan connected to the far end of the air duct entrained the ambient air to enhance heat dissipater. 2.1. Thermoelectric module It is noted that the effects of ceramic plated and joining copper traces and electrical contact resistances, which is small compared with the P-N type, are not considered at the back of the thermal
1 Q c ¼ STc I I2 R KðTh TC Þ 2
ð1Þ
1 Q h ¼ STh I þ I2 R KðTh TC Þ 2
ð2Þ
P ¼ Q h Q c ¼ SIðTh Tc Þ þ I2 R
ð3Þ
COP ¼
Qc P
ð4Þ
Setting dQ c/dI = 0 yields Imax as
Imax ¼
STc R
ð5Þ
The corresponding rate of maximum cooling (Q cmax) at the cold side is
Q cmax ¼
S2 T2C KðTh TC Þ 2R
ð6Þ
The value of Th-Tc at which Qc equal zero is defined as DTmax. It can be derived from Eq. (6).
DTmax ¼
S2 T2C 2RK
ð7Þ
In fact, the parameters of physical properties of a TEM are dependent on temperature. However, effect of temperature on the physical property is not in our consideration for simplifying assumption.
Y. Cai et al. / Energy Conversion and Management 124 (2016) 203–211
205
Fig. 1. Schematic diagram of the thermoelectric cooler.
temperature is related to the thermoelectric cold-side temperature by [20,25,26]
2.2. Analytical method As well known, the performance of heat dissipater is also significant for the cooling capacity and COP. The method of e NTU for the hot-side heat exchanger is employed to simplify the heat transfer module. The hot-side heat transfer supplied to TEC conversion process is defined by Hendricks and Karri [20] and Rowe [18]:
Qh ¼
Th Tin 1r ð1 rte Þ e Che;h þ Rjc;h
ð8Þ
Qh ¼
eh Ch ðTh Tin Þ ð1 rte Þð1 rhe;h Þ
ð9Þ
where is heat loss for the thermoelectric cooler, that is approximatively 10%. eh is the heat exchanger effectiveness. eh can be defined as
eh ¼ 1 exp
Uh Ah Ch
ð10Þ
ð14Þ
Lex Kex A
ð15Þ
Rex ¼
The electrical chip is replaced by micro-heater block made from aluminum corresponding to the real situation. Rint = 0.0054 W K1 for the contact resistance of interface between extender block to cold side of TEC or heat source. Where Tj is the surface temperature of electrical chip, Rex Kex Lex are the thermal resistance, thermal conductivity and length of extender block. and Rjc,c is the junction thermal resistance Rejc is equivalent cold side resistance. Finally, the comprehensive formula can be derived using Eqs. (1)–(15):
Qc ¼
½ðK þ SIÞuU h Ah S2 I2 Tj 12 I2 RðuU h Ah SI þ 2KÞ KuUh Ah Tin K SI þ uU h Ah þ ½ðK þ SIÞuU h Ah S2 I2 Rejc
ð16Þ
Combining Eqs. (9) and (10), the heat dissipating capacity is as:
Tj ¼
fK SI þ uU h Ah þ ½ðK þ SIÞuU h Ah S2 I2 Rejc gQ c þ 12 I2 RðuUh Ah SI þ 2KÞ þ KuUh Ah Tin ðK þ SIÞuU h Ah S2 I2
Q h ¼ uU h Ah ðTh Tin Þ
ð11Þ
where u is an equivalent thermal efficiency for hot-side heat exchanger, it can be defined as:
e
u¼
ln 11e
h
h ð1 rte Þð1 rhe;h Þ
ð12Þ
Reference from Hendricks [20,25], the heat transfer performance of hot-side can be represented by the value of UhAh, which is a fundamental parameter for heat exchanger. The UhAh performance can evaluate the effect of heat sink in thermoelectric heat exchanger mode. Present technologies indicted that finned heat pipe systems can achieve the high UhAh values. In this work, the UhAh value is considered for 40 W K1. Ch is 20 W K1 [21]. The cold side of thermoelectric cooler is connected with the electrical chip by aluminum extender block. Therefore, the chip
ð13Þ
Rjc;c ¼ Rex þ 2Rint
h h
Theoretically, the contact resistance Rjc;h between the base plate and the cold side of TEC is often small and negligible because of thermal grease, Thus, Qh is written as follows,
ðTj Tc Þ ð1 þ rte ÞðTj Tc Þ ¼ Rejc Rjc;c
Qc ¼
DT ¼
ð17Þ
ðSI uU h Ah ÞTj þ 12 I2 R½1 þ Rejc ð2SI uU h Ah Þ þ uU h Ah Tin ½1 þ SIRejc K SI þ uU h Ah þ Rejc ½ðK þ SIÞuU h Ah S2 I2
ð18Þ It is interesting to note that the above theoretical expression are derived in terms of the known ceiling temperature of electric devices Tj, the inlet temperature Tin and other parameters. The Tin can be controlled by air-conditioning system at 25 °C. The present analytical expressions can be compared with previous study and infrared measurement in order to be a reference for theory analysis.
3. Results and discussion Manufactures of TECs use the following parameters to specify their product: DTmax is the largest temperature differential that
206
Y. Cai et al. / Energy Conversion and Management 124 (2016) 203–211
can be obtained between the hot and cold ceramic plates at a given hot side temperature Tho, Imax is the input current which can produce the maximum DTmax across a TEC, Vmax is the DC voltage at the temperature difference of DTmax at I = Imax, Qmax is the maximum amount of heat that can be absorbed at the TEC cold side at I = Imax and DTmax = 0. The following equations based on the datasheet from manufactures can be used for calculating the parameters of the thermoelectric module [26,27],
R¼
ðTho DTmax ÞVmax Tho Imax
ð19Þ
K¼
ðTho DTmax ÞVmax Imax 2Tho DTmax
ð20Þ
S¼
Vmax Tho
ð21Þ
(a) Rex=0.03125Ω
Table 1 shows the characteristic data of TECs with various types. It is noting that the variation of thermoelectric properties is obviously different for S, K, R. According to the above equations, thermoelectric properties and the operating mode of TEM are studied. It assumes the ambient temperature is 25 °C which is easily controlled by Air conditioning room. 3.1. Effect of thermoelectric property Generally speaking, thermoelectric cooling system can be studied through the pellet thermoelectric parameters, and assessment by thermoelectric theory. However, for a fixed commercial product, the thermoelectric cooler which the geometric details always is not known. Therefore, the parameters of thermoelectric properties (S, R, K) can be obtained referring to the Eqs. (19)–(21) without resorting to the pellet thermoelectric parameters and geometric details. For the type B, C and D in Table 1, the dimensionless figure of merit ZT approximately is 0.77 at T = 300 K, which represents the efficiency of thermoelectric materials. The Qc and COP results at corresponding conditions versus the current I have been presented in Figs. 2 and 3, respectively. Both Rex = 0.03125(Lex = 0.01 m) and Rex = 0.15625 (Lex = 0.05 m) cases are considered. Effect of thermoelectric properties on three type thermoelectric coolers (B, C and D) are conducted in this TEC system. The plot for achieving the maximum cooling capacity and optimal COP at different current are shown in Fig. 2(a) for Rex = 0.03125 X, and Fig. 2(b) for (b) Rex = 0.15625 X. It is found that the cooling capacity increases with the current increases before I < Imax, and then decreases with the current increases after I > Imax. However, the COP decreases with the current increases. Compared with the Fig. 2(a) and (b). It is found that the Imax is diminished for the same type with increasing of the Rex. For the type D, the maximum cooling load is 99 W at Imax = 10.8 A in
(b) Rex=0.15625Ω Fig. 2. Plot of Qc and COP with I at different thermoelectric properties for (a) Rex = 0.03125 X (b) Rex = 0.15625 X, Tj = 50 °C.
Fig. 2(a), the same result obtain at Imax = 10.6 A in Fig. 2(b). The increasing of extender block thickness decrease the maximum cooling capacity for the same chip temperature (Tj = 50 °C). Mathematical calculation shows that high COP is achievable at Ish = 2.5A and the Qc = 99 W could approximately decrease 12% at same current. Unfortunately, the cooling capacity Qc is at low level when the COP is maximum. This phenomenon will be explained later. Instead of maximizing COP, a system optimal operating way shall be either to minimize Tj or to maximize Qc to fulfill the cooling task. The Tj and temperature difference (Th-Tc) with variation of current for two Rex values are shown in Fig. 3(a) and (b). Though the chip temperature is lower in Fig. 3(a) at the same current, the Imax also doesn’t change with scale of the extender block and so do temperature-difference (DT). On the other hand, changing Rex has little effect on the refrigerator performance for this case. As shown in Fig. 4, note that the COP with current at different Rex changes little. When the temperature difference is identical, the COP is also the same for a fixed cooling capacity. The thermoelectric properties have a great effect on the cooling load and chip
Table 1 List of different TECs with their supplied data and calculated module parameters. Number Type
A TB-199-2.0-0.9
B TB-127-2.0-1.15
C TEC2-12712
D TB-127-1.4-1.05
E TEC1-12706
Dimensions (mm) Imax (A) Qmax (W) Vmax (V) DTmax (K) Th0 (K) Z (1/K) S (V/K) K (W/K) R (X)
62 62 3.2 20.6 310 24.6 69 300 0.00259 0.082 2.8276 0.9195
55 55 3.4 17.6 156 15.7 69 300 0.00257 0.0523 1.5485 0.6869
40 40 3.2 10 113 15.2 69 300 0.00259 0.0507 0.8481 1.1704
40 40 3.3 8.6 84 15.7 69 300 0.00258 0.0523 0.7534 1.4057
40 40 3.9 6 56.5 15.2 68 300 0.00253 0.0507 0.5186 1.9591
Y. Cai et al. / Energy Conversion and Management 124 (2016) 203–211
207
(a) Rex=0.03125Ω (a) Rex=0.03125Ω
(b) Rex=0.15625Ω Fig. 3. Plot of Tj and DT with I at different thermoelectric properties for Qc = 80 W Rex = 0.03125 X (b) Rex = 0.15625 X.
temperature, which can be proved from Figs. 2–4. Though TEC B, C, D have a different maximum cooling power, Three type TEC show various thermal performance in the simulation calculation at Rex = 0.03125 X and 0.15625 X respectively. From the comparison, TEC B show major better thermal performance than TEC C, and the effect change bigger with the current increasing. Generally, the thermoelectric figure of merit Z indicates whether a thermoelectric material is a good for thermoelectric cooler. It depends on three material parameters: electrical resistance R, Seebeck coefficient S and thermal conductivity K. It is of interest to forecast the future predicted performance of TECs based on analysis method. If we assume all changes in Z lie in only one parameter S or K or R, and keep all other parameters unchangeable, we obtain the results indicated in Fig. 5. Clearly, the effect of Seebeck coefficient S states better thermal performance than that of K, R (S1 = 1.6S, K1 = 0.385 K, R1 = 0.384R at ZT = 2). Effects of parameters on TEM are various on minimizing Tj and maximizing Qc. For TEC B and D with the same Seebeck coefficient, a greater K and lower R result in a higher Qc. (From Fig. 5, Tj can deduce that K > K1, R > R1 and S > S1) This suggests that the differences of characteristic parameters exist obviously. It is achievable since the latest superlattices and quantum dots materials have the ZT as high as 2.4 at 300 K [28]. If parameters change for increasing ZT, TEC system is applicable and promising in cooling the electronics of the future. 3.2. The operating mode of TEM In order to investigate the operating mode of TEM for application in electronic devices, more operating conditions are simulated for the cooling capacity and COP. The relationship between cooling load and COP with current at Rex = 0 or Rex – 0 were investigated. As shown in Fig. 6, we should note that the cooling capacity and COP are relatively greater for Rex = 0 than that of the operating
(b) Rex=0.15625Ω Fig. 4. Plot of COP with I at different thermoelectric properties for (a) Rex = 0.03125 X (b) Rex = 0.15625 X, Qc = 80 W.
Fig. 5. Prediction of the future performance of thermoelectric coolers, assuming all changes in Z affect only S or K or R for ZT = 2 (or ZT = 1) (K > K1, R > R1 and S > S1).
condition with Rex – 0 at the same current, due to the fact that the extender block hinders the heat transfer. However, at the low current, the P is negative. This phenomenon can be illustrated by the following equations:
P ¼ SIðTh Tc Þ þ I2 R < 0 I < Ish ¼
SðTc Th Þ R
ð22Þ ð23Þ
208
Y. Cai et al. / Energy Conversion and Management 124 (2016) 203–211
Fig. 6. Plot of Qc and COP with I at with/without Rex for number A, Tj = 50 °C.
Fig. 8. Variation of temperature-difference DT with I for (a) Rex = 0 (b) Rex – 0.
Fig. 7. Plot of Qc and COP with I at different Tj for (a) Rex = 0 (b) Rex – 0, A.
So the temperature difference between heat source and the cold side of TEC is significant for the performance of TEC system. If the Rex = 0, the interface resistance is Rint. On the contrary, the Rex – 0, the interface resistance is 2Rint + Rex referring to Eq. (12). Within the scope of the design requirements, the extender block could be disadvantage for COP and Qc. Fig. 7 shows variation of the cooling capacity and COP with current at different chip temperature. Qc increases with the increase of the I before I < Imax, and then decreases after I > Imax. It is found that the maximum cooling capacity for Tj = 65 °C or Tj = 85 °C is over the maximum cooling capacity of thermoelectric module for type A is 310 W as shown in Table 1. Under this situation, the extender block can decrease the maximum cooling capacity for same chip temperature Tj. According to Fig. 7(a) and (b), Ish is lower for Rex – 0 compared with that of Rex = 0. This can demonstrate the extender
Fig. 9. Plot of Tj and COP with I at different Qc for (a) Rex = 0 (b) Rex – 0.
Y. Cai et al. / Energy Conversion and Management 124 (2016) 203–211
Fig. 10. Plot of Tj and COP with I at different Qc for (a) Rex = 0 (b) Rex – 0.
block can decrease COP for same chip temperature. The corresponding temperature differences imply that I0 in Fig. 8(a) is comparatively greater compared with that in Fig. 8(b) and the operating mode change with current. Previous studies of thermoelectric system has examined that the maximum cooling load is under the DT = 0, but it is not the case. The variation of the chip temperature Tj and COP with working electrical current under different cooling capacity are analyzed, as shown in Fig. 9. Compared Fig. 9(a) and (b), the lower Tj can be obtained at Rex = 0 for the same cooling capacity. It is interesting that the Ish changes a little with/without Rex. Fig. 10(a) and (b) shows the variation of temperature difference with current at Rex = 0 or Rex – 0. We can find that equilibrium value In is identical counterpart for Fig. 10(a) and (b). For the fixed current, we can obtain an expression for temperature difference:
Tc Th ¼
In R S
209
It is interesting to note that the current (In) can be analogous by setting Tc Th = 0. For the known cooling capacity, the variation of the optimal current (Imax) under Rex = 0 or Rex – 0 are not very large, according to Fig. 9(a) and (b). Certainly, the optimal current for minimizing Tj at Qc = 80 W is 16.2–17.2 A, however, Imax is 17.2–17.6A at Qc = 120 W in Fig. 9(a). Also the neutral current In in temperature-difference has slightly increased corresponding to Rex = 0 or Rex – 0 in Fig. 10(a) and (b). To illustrate the applications of the framework developed above for the analysis of TEMs, the parameters governing refrigeration mode are developed. Table 2 may be used to determine the operating mode of TEM system from Qc, Tc Th, P and I. When 0 < Tc Th < DTmax,h, the variation of current is at range 0 from In. In < Imax in Figs. 7 and 8, however, In > Imax in Fig. 11. Assuming In > Imax, as shown in Fig. 11, the TEM operates in generation mode due to P < 0. At I = Ish, the TEM operates in short circuit cooling mode. At Ish < I < Imax and Imax < I < In, the TEM operates in efficient and inefficient cooling modes, respectively. When 0 < Th Tc < DTmax,c. The variation of current is at range from In to Isat. In this aspect, for In < I < Ii, a TEM operates in Inefficient cooling mode. Cooling mode exits for Ish < I < Ii because of Qc is positive bigger than Qc,open. Qc,open is cooling quantity under I = 0. Ii is defined as the current for which Qc equals its open circuit (I = 0 as Qc = Q c,open). So Ii = 2Imax. Certainly, ICOP=0.5 can be as a reference for evaluating the performance of TEC system. The optimal operating mode of TEM for cooling mode is at Ish < I < ICOP=0.5. It is defined that ICOP=0.5 is a threshold value. The major disadvantage of TEC system is that their thermodynamic efficiencies are smaller than those of competing technologies, such as vapor compression refrigerators. This is exacerbated by the fact that TEM normally require or produce low voltage/high current DC power because of the losses associated with voltage conversion. So it is beneficial for designers making operating mode of TEM towards a favorable direction. As shown in the Table 2, an efficient cooling operating mode could be acceptable in order to maintain the temperature of the object or taking away the cooling
ð24Þ
Table 2 Operating mode criteria and relationship of parameters in thermoelectric cooling system. Primary mode
Sub-mode
Qc
Temperaturedifference
P
I
Generation Neutral Cooling Cooling Neutral Cooling Heating Heating
+I generation Short circuit Efficient Inefficient – Inefficient – –
>0 >0 >0 >0 >0 >0 >0 >0 or < 0
Tc > T h Tc > Th Tc > Th Tc > Th Tc = Th Tc < Th Tc < Th Th > Tc or Th-Tc > DTmax
<0 =0 >0 >0 >0 >0 >0 >0
0 < I < Ish I = Ish Ish < I < Imax Imax < I < In I = In In < I < Ii Ii < I < Isat I > Isat Fig. 11. Variation of Qc, DT and COP with I at with extender block for Tj = 50 °C, E.
210
Y. Cai et al. / Energy Conversion and Management 124 (2016) 203–211
Fig. 12. Comparison between present analytical results and literature with different current.
load at required. Conditions parameters corresponding to minimize Tj or maximize Qc are considered though all conditions can’t reach at the same time.
Fig. 13. Comparison between present analytical results and measure with different current (E).
TEM module is considered from the parameters. It is seen that the Tj is obviously lower at L = 1 cm, which Rex is 0.03125 X. Also the analysis is agree with the measurement at 2–4 A. 4. Conclusions
3.3. Comparison with previous studies The present analysis method is compared with previous studies based on the two parts, one is for the thermoelectric properties, and the other is for chip temperature. Firstly, the performance is enhanced due to the effect of Seebeck coefficient with increasing of ZT has been presented in reference [29]. An assuming future predicted performance of TECs was discussed, which has the effect of increasing COP dramatically. For example, some detailed parameters for ZT = 2 on account of changing in Seebeck coefficient S can be found. The interesting temperature point of Tj = 50 °C at 170 W and Tj = 85 °C at 217 W are obtained from Fig. 5, greater than which corresponding result is 128 W and 169 W in the literature [29]. It is about 30% rise for improving cooling capacity. Another comparison based on the study of Zhang [13,14] is presented to verify the reliability of the present work. In their work, the expression of chip temperature was derived under the thermoelectric properties. The experimental study was conducted based on a processor test platform. In order to verify the correctness of the current research methods for operating mode of TEM system. Some parameters covering with Rjc,c and uU h Ah and Ta should be set to the same as the references. The comparison results for reference are shown in Fig. 12. It is found that the Tj under Qc = 100 W or 140 W is fitting with literature. This indicates the method of e NTU applied in electrical devices in this work is reliable and acceptable for predicting future TEC system. Fig. 12 also shows a comparison between measured Qc by Zhang with analytical results at Tj = 25 °C. Clearly, the present analysis is according with the experimental measure within a certain error range.
Based on the thermoelectric theory and the method of e NTU, a novel operating mode of thermoelectric module are established for electronic devices cooling system. The main feature of the present analytical method is that it makes full use of TEM parameters and studies the scale of extender block under two set analytical condition, that is, the chip temperature Tj at a fixed cooling capacity Qc and Qc at a fixed Tj, respectively. Theoretical expression of Tj, Qc and DT are derived in terms of present analytical method. Thermoelectric cooling system has been studied based on the thermoelectric module properties, and the optimal current for maximum cooling capacity and minimum chip temperature can be obtained under corresponding fixed condition, such as Rex. Results have demonstrated that thermoelectric properties make a great effect on the cooling capacity and chip temperature. The extender block shows better thermal performance in cooling capacity than chip temperature, so do COP and temperature difference. Future predict due to enhancing ZT only by changing Seebeck coefficient indicates Seebeck coefficient can improve cooling capacity compared with electrical resistance and thermal conductivity. The operating modes of thermoelectric module covering with control parameter are developed for getting a favorable cooling mode in this analytical method. The COP and temperature difference are as a reference for cooling mode forecasting. The cooling mode of TEM are within Ish < Ii. Finally, the validation of the present analysis is also conducted compared with previous studies and through the infrared thermal imager. Measuring results suggest good agreement is achieved between measurement and analysis.
3.4. Experiment validation Acknowledgements The experimental measurement was conducted by corresponding experiment table, which the heat source was replaced. A full scale experiment is setting up on the account of schematic diagram of thermoelectric cooler as shown in the Fig. 1. Two set measure under L = 1 cm and L = 5 cm are presented for comparison with temperature distribution to record Tj. Given the limited space available, TEC current and Qc are fixed to obtained temperature. TEC 12706 was selected as the test. Fig. 13 shows a surface temperature of heat source comparison of L = 1 cm and L = 5 cm at specified Qc = 45 W. As previous analysis, the operating cooling mode of
Authors would gratefully acknowledge the financial support of National Natural Science Foundation of China, (NSFC, Grant No. 51304233, Multi-physics diffusion of leaked natural gas and backward time inverse identification of leakage sources; NSFC, Grant No. 51208192, Instability theory and inverse convection design of air flow patterns in the large space), Hong Kong Scholar Program (XJ2013042), China Postdoctoral Science Foundation (2014M560593), Qingdao Postdoctoral Science Foundation, Fundamental Research Funding Program for National Key Universities in
Y. Cai et al. / Energy Conversion and Management 124 (2016) 203–211
China, Postgraduate Innovation Program (YCXJ2016073). In addition, Fu-Yun Zhao would like to acknowledge the financial support of the Thousand Youth Talents Plan from the Organization Department of CCP Central Committee (Wuhan University, China, Grant No. 208273259), Hunan Provincial Natural Science Foundation for Distinguished Young Scholars supported by Hunan Provincial Government (Grant No. 14JJ1002, Multiple fluid mechanisms of urban ventilation and its safety through source identification), Hubei Provincial Natural Science Foundation of Hubei Province (Grant No. 2015CFB261, Multiple macro flow states of urban built environment), Scientific Research Foundation for the Returned Overseas Chinese Scholars, State Education Ministry (Grant No. 230303, inverse identification of unperceived hazardous sources coupling with organized safe evacuation in the public buildings), National Key Basic Research Program of China (973 Program, Grant No. 2014CB239203), and Twelve-Five National Supportive Plan from Ministry of Science and Technology of China (Grant No. 2011BAJ03B07). References [1] Hanreich G, Nicolics J, Musiejovsky L. High resolution thermal simulation of electronic components. Microelectron Reliab 2000;40:2069–76. [2] Bar-Cohen A, Kraus AD, Davidson SF. Thermal frontiers in the design and packaging microelectronic equipment. Mech Eng 1983;105:53–9. [3] Zhang HY, Pinjala D, Teo PS. Thermal management of high power dissipation electronic packages: from air cooling to liquid cooling. In: IEEE, 5th electronics packaging technology conference (EPTC). p. 620–5. [4] Chein R, Huang G. Thermoelectric cooler application in electronic cooling. Appl Therm Eng 2004;24:2207–17. [5] Wang J, Zhao X-J, Cai Y-X, Zhang C, Bao W-W. Experimental study on the thermal management of high-power LED headlight cooling device integrated with thermoelectric cooler package. Energy Convers Manage 2015;101:532–40. [6] Xu D. Thermoelectric refrigeration and applied technology. Shanghai: Shanghai Jiaotong University Press; 1992. [7] Hodes M. On one-dimensional analysis of thermoelectric modules (TEMs). IEEE Trans Comp Packag Technol 2005;28:218–29. [8] Huang BJ, Chin CJ, Duang CL. A design method of thermoelectric cooler. Int J Refrig 2000;23(3):208–18. [9] Chen J, Yu J, Ma M. Theoretical study on an integrated two-stage cascaded thermoelectric module operating with dual power sources. Energy Convers Manage 2015;98:28–33.
211
[10] Lineykin S, Ben-Yaakov S. User-friendly and intuitive graphical approach to the design of thermoelectric cooling systems. Int J Refrig 2007;30(5):798–804. [11] Manikandan S, Kaushik SC. Energy and exergy analysis of an annular thermoelectric cooler. Energy Convers Manage 2015;106:804–14. [12] Simons RE, Ellsworth MJ, Chu RC. An assessment of module cooling enhancement with thermoelectric coolers. J Heat Transfer-Trans ASME 2005;127(1):76–84. [13] Zhang HY. A general approach in evaluating and optimizing thermoelectric coolers. Int J Refrig 2010;33:1187–96. [14] Zhang HY, Mui YC, Tarin M. Analysis of thermoelectric cooler performance for high power electronic packages. Appl Therm Eng 2010;30:561–8. [15] Liu Di, Zhao Fu-Yun, Yang Hong-Xing, Tang Guang-Fa. Thermoelectric mini cooler coupled with micro thermosiphon for CPU cooling system. Energy 2015;83:29–36. [16] Liu D, Zhao FY, Tang GF. Modeling and performance investigation of a closedtype thermoelectric clothes dryer. Dry Technol 2008;26:1208–16. [17] Bergman TL, Inreopera FP, lavine AS, Dewitt DP. Fundamentals of heat and mass transfer. 7th ed. John Wiley & Sons; 2011. [18] Rowe DM. Modules, systems and applications in thermoelectrics. CRC Press; 2011. [19] Crane DT, Bell LE. Design to maximize performance of a thermoelectric power generator with a dynamic thermal power source. J Energy Res Technol 2009;131:012401-1–1-8. [20] Hendricks T, Karri N. Micro- and nano-technology: a critical design key in advanced thermoelectric cooling systems. J Electron Mater 2009;38 (7):1257–67. [21] Zhou Y, Yu J. Design optimization of thermoelectric cooling systems for applications in electronic devices. Int J Refrig 2012;35(4):1139–44. [22] Meng JH, Wang XD, Zhang XX. Transient modeling and dynamic characteristics of thermoelectric cooler. Appl Energy 2013;108:340–8. [23] Zhao D, Tan G. A review of thermoelectric cooling: materials, modeling and applications. Appl Therm Eng 2014;66:15–24. [24] Seifert W, Ueltzen M, Müller E. One-dimensional modelling of thermoelectric cooling. Phys Status Solidi A 2002;194(1):277–90. [25] Hendricks T. Thermal system interactions in optimizing advanced thermoelectric energy recovery systems. J Energy Resour Technol 2006;129:223–31. [26] Luo Z. A simple method to estimate the physical characteristics of a thermoelectric cooler from vendor datasheets. Electron Cool 2008;14:22–7. [27] Lineykin S, Ben-Yaakov S. Analysis of thermoelectric coolers by a spicecompatible equivalent 0 circuit model. IEEE Power Electron Lett 2005;3:63–6. [28] Elsheikh MH, Shnawah DA, Sabri MFM, Said SBM, Hassan MH, Bashir MBA, et al. A review on thermoelectric renewable energy: principle parameters that affect their performance. Renew Sustain Energy Rev 2014;30:337–55. [29] Phelan PE, Chiriac VA, Lee TY. Current and future miniature refrigeration cooling technologies for high power microelectronics. IEEE Trans Comp Pack Technol 2002;25:356–65.
Contents lists available at ScienceDirect
Energy Conversion and Management journal homepage: www.elsevier.com/locate/enconman
Performance analysis and assessment of thermoelectric micro cooler for electronic devices Yang Cai a, Di Liu a,⇑, Fu-Yun Zhao b, Jian-Feng Tang a a b
College of Pipeline and Civil Engineering, China University of Petroleum, 266580 Qingdao, Shandong Province, China School of Power and Mechanical Engineering, Wuhan University, Wuhan, Hubei Province, China
a r t i c l e
i n f o
Article history: Received 25 February 2016 Received in revised form 29 June 2016 Accepted 3 July 2016
Keywords: Thermoelectric cooling system COP e NTU Operating mode
a b s t r a c t A novel operating mode of thermoelectric module (TEM, cooling, heating, generation) is established for electronic devices cooling, based on the method of effectiveness-number of transfer units (e NTUÞ. This work mainly focused on the effect of thermoelectric properties and the scale of extender block on cooling performance under different operating conditions in order to obtain effective cooling operating mode. Based on the TEM parameters, two sets of analytical solutions for thermoelectric cooler (TEC) are derived for the chip temperature Tj at a fixed cooling power Qc and Qc at a fixed Tj, respectively. The performance of TEC with/without scale of extender block is studied for the lowest chip temperature and maximum cooling capacity at fixed conditions. Analysis results show the thermoelectric properties and extender block are significant characteristics for different operating conditions. However, the coefficient of performance (COP) and temperature difference changed a little under given thermoelectric properties. The results indicate TEC system applied in electronic devices obtains effectively cooling module by controlling operating parameters, which do not changed with scale of extender block. The validation of the present analysis is also conducted compared with previous studies and through the infrared thermal imager. Ó 2016 Elsevier Ltd. All rights reserved.
1. Introduction High power electronic devices with smaller size and higher encapsulation have resulted in the focus on the thermal management [1]. Investigations show the thermal reliability of a silicon chip decreases by about 10% for a rise of 2 °C in temperature [2], resulting in the increasing of thermal management challenges. Simultaneously, traditional passive cooling technology such as natural convection cooling, forced air cooling, liquid cooling, are reaching the limits of cooling capacity and cooling efficiency for high power electric devices [3–5]. Compared with traditional cooling, TEC is thought to be one of the preferable technologies due to its numerous advantages, such as small volume, environmentally friendly and temperature control capability [6]. Therefore, developing thermoelectric micro cooler system has potential to make important contributions to the growing thermal management challenges of electronic devices. Most of the previous work on the TEC system has examined optimization and performance improvement to electronic devices by building various thermoelectric modules [7–16]. Zhang [13] ⇑ Corresponding author. E-mail addresses: [email protected] (D. Liu), [email protected] (F.-Y. Zhao). http://dx.doi.org/10.1016/j.enconman.2016.07.011 0196-8904/Ó 2016 Elsevier Ltd. All rights reserved.
and Zhang et al. [14] presented a generalized theoretical model in evaluating and optimizing thermoelectric coolers. Their modules did not consider the heat exchange in heat sink through heat exchange medium. Thermal design method covering with e NTU for solving the heat dissipater problem of TEC system have been investigated [17–21]. It is noted that their module expressions were not involved the temperature-difference and operating mode of effective cooling module. At present, the operating mode of thermoelectric module in practical use have not developed because of the dynamic process covering with temperature difference with current. Only Hodes [7] investigated the operating mode of TEM under the operating parameters to illustrate the application of the analysis framework. However, for a fixed hot and cold side, the energy conversion is changing with the temperature difference, and the analysis is extended to practical use, such as electronic heat dissipater. To the best of author’ knowledge, most of previous studies are heat dissipater for electronic devices focused on thermoelectric configuration and optimizing parameters. The thermal design theory (e NTU) applied in thermoelectric cooling system have not been developed. This work focuses on the operating mode of TEM under the different scale of extender block for various thermoelectric properties. Based on the thermal design theory, the
204
Y. Cai et al. / Energy Conversion and Management 124 (2016) 203–211
Nomenclature A C COP I Ii In Isat Ish K Kex L Q Qc,open Qcmax R S T Tho Tj DT DT max U UA Z ZT
area (m2) specific heat (W K1) coefficient of performance current (A) current at open circuit (I = 0) value (Qc,open) (A) neutral current at DT ¼ 0 (A) current at Qc = 0 (A) short-circuit current (A) thermal conductance (W K1) thermal conductance of extender block (W K1) length (m) heat flux (W) cooling quantity with I = 0 (W) maximum cooling load (W) resistance (X) Seebeck coefficient (V K1) temperature (K) hot side reference temperature (K) surface temperature of chip (K) TEC hot side to cold side temperature difference (K) maximum temperature-difference(K) overall heat transfer coefficient (W m2 K1) heat performance parameter (W K1) figure of merit (1/K) dimensionless figure of merit
Greek symbols r heat loss efficient rhe;h heat loss efficient of heat exchanger rte heat loss efficient of thermoelectric cooler e heat exchanger efficient u equivalent thermal efficiency Subscript c ejc ex h he in jc jc,c jc,h max n sh te
TEC cold side equivalent cold side junction to TEC extender block TEC hot side hot-side heat exchanger inlet junction to TEC cold side junction to TEC hot side junction to TEC Maximum Neutral short-circuit thermoelectric cooler
expression of cooling power and chip temperature are obtained for considering various operating mode. Two sets of analytical solutions for thermoelectric cooler are derived for the chip temperature at a fixed cooling power and cooling power at a fixed chip temperature, respectively. The objective of operating mode of TEM is to establish effective cooling module to predict the efficiency of TEC system and set up references for TEC system. Comparison with previous work and the full-scale experimental are also conducted for verification of the present analysis method.
balance equation. Thus, the theoretical equation for the TEM can be given as follows [23,24],
2. Schematic diagram of TEC system
where I is the electric current, Th and Tc are the junction temperature of thermoelectric cooler. The Seekbeck coefficient S, thermal conductivity K and the total electrical resistance R are the physical properties of a TEM, which is important for the thermoelectric module. Thermoelectric cooler COP is defined as
As shown in Fig. 1, the schematic of TEC system consists of TEM, hot side heat exchanger, extender block, and heater block. The TEM comprises many pairs of P-N type semiconductor columns, metallic connectors, and two electrically insulting ceramic plates [22]. The interface face contacted with electrical device by extender block is the cold side of thermoelectric cooler. Inversely, the interface face connected with heat sink is the hot side of thermoelectric cooler, and in this paper the heat sink with fins is called as hot side heat exchanger. The extender block, which could avert decreasing the cooling capacity and COP, has been considered as a buffer plate. A hot side heat exchanger with fin is used to take away the heat from heat source in the hot-side of TEC. The design task is to minimize the chip temperature or maximum cooling thermoelectric cooling load through TEC enhanced air cooling or others cooling ways. It is noted that the hot or cold side of TEC is not dependent on the temperature level. The contact of thermoelectric two side and aluminum extender is coated with thermal grease for decreasing thermal resistance. An axial fan connected to the far end of the air duct entrained the ambient air to enhance heat dissipater. 2.1. Thermoelectric module It is noted that the effects of ceramic plated and joining copper traces and electrical contact resistances, which is small compared with the P-N type, are not considered at the back of the thermal
1 Q c ¼ STc I I2 R KðTh TC Þ 2
ð1Þ
1 Q h ¼ STh I þ I2 R KðTh TC Þ 2
ð2Þ
P ¼ Q h Q c ¼ SIðTh Tc Þ þ I2 R
ð3Þ
COP ¼
Qc P
ð4Þ
Setting dQ c/dI = 0 yields Imax as
Imax ¼
STc R
ð5Þ
The corresponding rate of maximum cooling (Q cmax) at the cold side is
Q cmax ¼
S2 T2C KðTh TC Þ 2R
ð6Þ
The value of Th-Tc at which Qc equal zero is defined as DTmax. It can be derived from Eq. (6).
DTmax ¼
S2 T2C 2RK
ð7Þ
In fact, the parameters of physical properties of a TEM are dependent on temperature. However, effect of temperature on the physical property is not in our consideration for simplifying assumption.
Y. Cai et al. / Energy Conversion and Management 124 (2016) 203–211
205
Fig. 1. Schematic diagram of the thermoelectric cooler.
temperature is related to the thermoelectric cold-side temperature by [20,25,26]
2.2. Analytical method As well known, the performance of heat dissipater is also significant for the cooling capacity and COP. The method of e NTU for the hot-side heat exchanger is employed to simplify the heat transfer module. The hot-side heat transfer supplied to TEC conversion process is defined by Hendricks and Karri [20] and Rowe [18]:
Qh ¼
Th Tin 1r ð1 rte Þ e Che;h þ Rjc;h
ð8Þ
Qh ¼
eh Ch ðTh Tin Þ ð1 rte Þð1 rhe;h Þ
ð9Þ
where is heat loss for the thermoelectric cooler, that is approximatively 10%. eh is the heat exchanger effectiveness. eh can be defined as
eh ¼ 1 exp
Uh Ah Ch
ð10Þ
ð14Þ
Lex Kex A
ð15Þ
Rex ¼
The electrical chip is replaced by micro-heater block made from aluminum corresponding to the real situation. Rint = 0.0054 W K1 for the contact resistance of interface between extender block to cold side of TEC or heat source. Where Tj is the surface temperature of electrical chip, Rex Kex Lex are the thermal resistance, thermal conductivity and length of extender block. and Rjc,c is the junction thermal resistance Rejc is equivalent cold side resistance. Finally, the comprehensive formula can be derived using Eqs. (1)–(15):
Qc ¼
½ðK þ SIÞuU h Ah S2 I2 Tj 12 I2 RðuU h Ah SI þ 2KÞ KuUh Ah Tin K SI þ uU h Ah þ ½ðK þ SIÞuU h Ah S2 I2 Rejc
ð16Þ
Combining Eqs. (9) and (10), the heat dissipating capacity is as:
Tj ¼
fK SI þ uU h Ah þ ½ðK þ SIÞuU h Ah S2 I2 Rejc gQ c þ 12 I2 RðuUh Ah SI þ 2KÞ þ KuUh Ah Tin ðK þ SIÞuU h Ah S2 I2
Q h ¼ uU h Ah ðTh Tin Þ
ð11Þ
where u is an equivalent thermal efficiency for hot-side heat exchanger, it can be defined as:
e
u¼
ln 11e
h
h ð1 rte Þð1 rhe;h Þ
ð12Þ
Reference from Hendricks [20,25], the heat transfer performance of hot-side can be represented by the value of UhAh, which is a fundamental parameter for heat exchanger. The UhAh performance can evaluate the effect of heat sink in thermoelectric heat exchanger mode. Present technologies indicted that finned heat pipe systems can achieve the high UhAh values. In this work, the UhAh value is considered for 40 W K1. Ch is 20 W K1 [21]. The cold side of thermoelectric cooler is connected with the electrical chip by aluminum extender block. Therefore, the chip
ð13Þ
Rjc;c ¼ Rex þ 2Rint
h h
Theoretically, the contact resistance Rjc;h between the base plate and the cold side of TEC is often small and negligible because of thermal grease, Thus, Qh is written as follows,
ðTj Tc Þ ð1 þ rte ÞðTj Tc Þ ¼ Rejc Rjc;c
Qc ¼
DT ¼
ð17Þ
ðSI uU h Ah ÞTj þ 12 I2 R½1 þ Rejc ð2SI uU h Ah Þ þ uU h Ah Tin ½1 þ SIRejc K SI þ uU h Ah þ Rejc ½ðK þ SIÞuU h Ah S2 I2
ð18Þ It is interesting to note that the above theoretical expression are derived in terms of the known ceiling temperature of electric devices Tj, the inlet temperature Tin and other parameters. The Tin can be controlled by air-conditioning system at 25 °C. The present analytical expressions can be compared with previous study and infrared measurement in order to be a reference for theory analysis.
3. Results and discussion Manufactures of TECs use the following parameters to specify their product: DTmax is the largest temperature differential that
206
Y. Cai et al. / Energy Conversion and Management 124 (2016) 203–211
can be obtained between the hot and cold ceramic plates at a given hot side temperature Tho, Imax is the input current which can produce the maximum DTmax across a TEC, Vmax is the DC voltage at the temperature difference of DTmax at I = Imax, Qmax is the maximum amount of heat that can be absorbed at the TEC cold side at I = Imax and DTmax = 0. The following equations based on the datasheet from manufactures can be used for calculating the parameters of the thermoelectric module [26,27],
R¼
ðTho DTmax ÞVmax Tho Imax
ð19Þ
K¼
ðTho DTmax ÞVmax Imax 2Tho DTmax
ð20Þ
S¼
Vmax Tho
ð21Þ
(a) Rex=0.03125Ω
Table 1 shows the characteristic data of TECs with various types. It is noting that the variation of thermoelectric properties is obviously different for S, K, R. According to the above equations, thermoelectric properties and the operating mode of TEM are studied. It assumes the ambient temperature is 25 °C which is easily controlled by Air conditioning room. 3.1. Effect of thermoelectric property Generally speaking, thermoelectric cooling system can be studied through the pellet thermoelectric parameters, and assessment by thermoelectric theory. However, for a fixed commercial product, the thermoelectric cooler which the geometric details always is not known. Therefore, the parameters of thermoelectric properties (S, R, K) can be obtained referring to the Eqs. (19)–(21) without resorting to the pellet thermoelectric parameters and geometric details. For the type B, C and D in Table 1, the dimensionless figure of merit ZT approximately is 0.77 at T = 300 K, which represents the efficiency of thermoelectric materials. The Qc and COP results at corresponding conditions versus the current I have been presented in Figs. 2 and 3, respectively. Both Rex = 0.03125(Lex = 0.01 m) and Rex = 0.15625 (Lex = 0.05 m) cases are considered. Effect of thermoelectric properties on three type thermoelectric coolers (B, C and D) are conducted in this TEC system. The plot for achieving the maximum cooling capacity and optimal COP at different current are shown in Fig. 2(a) for Rex = 0.03125 X, and Fig. 2(b) for (b) Rex = 0.15625 X. It is found that the cooling capacity increases with the current increases before I < Imax, and then decreases with the current increases after I > Imax. However, the COP decreases with the current increases. Compared with the Fig. 2(a) and (b). It is found that the Imax is diminished for the same type with increasing of the Rex. For the type D, the maximum cooling load is 99 W at Imax = 10.8 A in
(b) Rex=0.15625Ω Fig. 2. Plot of Qc and COP with I at different thermoelectric properties for (a) Rex = 0.03125 X (b) Rex = 0.15625 X, Tj = 50 °C.
Fig. 2(a), the same result obtain at Imax = 10.6 A in Fig. 2(b). The increasing of extender block thickness decrease the maximum cooling capacity for the same chip temperature (Tj = 50 °C). Mathematical calculation shows that high COP is achievable at Ish = 2.5A and the Qc = 99 W could approximately decrease 12% at same current. Unfortunately, the cooling capacity Qc is at low level when the COP is maximum. This phenomenon will be explained later. Instead of maximizing COP, a system optimal operating way shall be either to minimize Tj or to maximize Qc to fulfill the cooling task. The Tj and temperature difference (Th-Tc) with variation of current for two Rex values are shown in Fig. 3(a) and (b). Though the chip temperature is lower in Fig. 3(a) at the same current, the Imax also doesn’t change with scale of the extender block and so do temperature-difference (DT). On the other hand, changing Rex has little effect on the refrigerator performance for this case. As shown in Fig. 4, note that the COP with current at different Rex changes little. When the temperature difference is identical, the COP is also the same for a fixed cooling capacity. The thermoelectric properties have a great effect on the cooling load and chip
Table 1 List of different TECs with their supplied data and calculated module parameters. Number Type
A TB-199-2.0-0.9
B TB-127-2.0-1.15
C TEC2-12712
D TB-127-1.4-1.05
E TEC1-12706
Dimensions (mm) Imax (A) Qmax (W) Vmax (V) DTmax (K) Th0 (K) Z (1/K) S (V/K) K (W/K) R (X)
62 62 3.2 20.6 310 24.6 69 300 0.00259 0.082 2.8276 0.9195
55 55 3.4 17.6 156 15.7 69 300 0.00257 0.0523 1.5485 0.6869
40 40 3.2 10 113 15.2 69 300 0.00259 0.0507 0.8481 1.1704
40 40 3.3 8.6 84 15.7 69 300 0.00258 0.0523 0.7534 1.4057
40 40 3.9 6 56.5 15.2 68 300 0.00253 0.0507 0.5186 1.9591
Y. Cai et al. / Energy Conversion and Management 124 (2016) 203–211
207
(a) Rex=0.03125Ω (a) Rex=0.03125Ω
(b) Rex=0.15625Ω Fig. 3. Plot of Tj and DT with I at different thermoelectric properties for Qc = 80 W Rex = 0.03125 X (b) Rex = 0.15625 X.
temperature, which can be proved from Figs. 2–4. Though TEC B, C, D have a different maximum cooling power, Three type TEC show various thermal performance in the simulation calculation at Rex = 0.03125 X and 0.15625 X respectively. From the comparison, TEC B show major better thermal performance than TEC C, and the effect change bigger with the current increasing. Generally, the thermoelectric figure of merit Z indicates whether a thermoelectric material is a good for thermoelectric cooler. It depends on three material parameters: electrical resistance R, Seebeck coefficient S and thermal conductivity K. It is of interest to forecast the future predicted performance of TECs based on analysis method. If we assume all changes in Z lie in only one parameter S or K or R, and keep all other parameters unchangeable, we obtain the results indicated in Fig. 5. Clearly, the effect of Seebeck coefficient S states better thermal performance than that of K, R (S1 = 1.6S, K1 = 0.385 K, R1 = 0.384R at ZT = 2). Effects of parameters on TEM are various on minimizing Tj and maximizing Qc. For TEC B and D with the same Seebeck coefficient, a greater K and lower R result in a higher Qc. (From Fig. 5, Tj can deduce that K > K1, R > R1 and S > S1) This suggests that the differences of characteristic parameters exist obviously. It is achievable since the latest superlattices and quantum dots materials have the ZT as high as 2.4 at 300 K [28]. If parameters change for increasing ZT, TEC system is applicable and promising in cooling the electronics of the future. 3.2. The operating mode of TEM In order to investigate the operating mode of TEM for application in electronic devices, more operating conditions are simulated for the cooling capacity and COP. The relationship between cooling load and COP with current at Rex = 0 or Rex – 0 were investigated. As shown in Fig. 6, we should note that the cooling capacity and COP are relatively greater for Rex = 0 than that of the operating
(b) Rex=0.15625Ω Fig. 4. Plot of COP with I at different thermoelectric properties for (a) Rex = 0.03125 X (b) Rex = 0.15625 X, Qc = 80 W.
Fig. 5. Prediction of the future performance of thermoelectric coolers, assuming all changes in Z affect only S or K or R for ZT = 2 (or ZT = 1) (K > K1, R > R1 and S > S1).
condition with Rex – 0 at the same current, due to the fact that the extender block hinders the heat transfer. However, at the low current, the P is negative. This phenomenon can be illustrated by the following equations:
P ¼ SIðTh Tc Þ þ I2 R < 0 I < Ish ¼
SðTc Th Þ R
ð22Þ ð23Þ
208
Y. Cai et al. / Energy Conversion and Management 124 (2016) 203–211
Fig. 6. Plot of Qc and COP with I at with/without Rex for number A, Tj = 50 °C.
Fig. 8. Variation of temperature-difference DT with I for (a) Rex = 0 (b) Rex – 0.
Fig. 7. Plot of Qc and COP with I at different Tj for (a) Rex = 0 (b) Rex – 0, A.
So the temperature difference between heat source and the cold side of TEC is significant for the performance of TEC system. If the Rex = 0, the interface resistance is Rint. On the contrary, the Rex – 0, the interface resistance is 2Rint + Rex referring to Eq. (12). Within the scope of the design requirements, the extender block could be disadvantage for COP and Qc. Fig. 7 shows variation of the cooling capacity and COP with current at different chip temperature. Qc increases with the increase of the I before I < Imax, and then decreases after I > Imax. It is found that the maximum cooling capacity for Tj = 65 °C or Tj = 85 °C is over the maximum cooling capacity of thermoelectric module for type A is 310 W as shown in Table 1. Under this situation, the extender block can decrease the maximum cooling capacity for same chip temperature Tj. According to Fig. 7(a) and (b), Ish is lower for Rex – 0 compared with that of Rex = 0. This can demonstrate the extender
Fig. 9. Plot of Tj and COP with I at different Qc for (a) Rex = 0 (b) Rex – 0.
Y. Cai et al. / Energy Conversion and Management 124 (2016) 203–211
Fig. 10. Plot of Tj and COP with I at different Qc for (a) Rex = 0 (b) Rex – 0.
block can decrease COP for same chip temperature. The corresponding temperature differences imply that I0 in Fig. 8(a) is comparatively greater compared with that in Fig. 8(b) and the operating mode change with current. Previous studies of thermoelectric system has examined that the maximum cooling load is under the DT = 0, but it is not the case. The variation of the chip temperature Tj and COP with working electrical current under different cooling capacity are analyzed, as shown in Fig. 9. Compared Fig. 9(a) and (b), the lower Tj can be obtained at Rex = 0 for the same cooling capacity. It is interesting that the Ish changes a little with/without Rex. Fig. 10(a) and (b) shows the variation of temperature difference with current at Rex = 0 or Rex – 0. We can find that equilibrium value In is identical counterpart for Fig. 10(a) and (b). For the fixed current, we can obtain an expression for temperature difference:
Tc Th ¼
In R S
209
It is interesting to note that the current (In) can be analogous by setting Tc Th = 0. For the known cooling capacity, the variation of the optimal current (Imax) under Rex = 0 or Rex – 0 are not very large, according to Fig. 9(a) and (b). Certainly, the optimal current for minimizing Tj at Qc = 80 W is 16.2–17.2 A, however, Imax is 17.2–17.6A at Qc = 120 W in Fig. 9(a). Also the neutral current In in temperature-difference has slightly increased corresponding to Rex = 0 or Rex – 0 in Fig. 10(a) and (b). To illustrate the applications of the framework developed above for the analysis of TEMs, the parameters governing refrigeration mode are developed. Table 2 may be used to determine the operating mode of TEM system from Qc, Tc Th, P and I. When 0 < Tc Th < DTmax,h, the variation of current is at range 0 from In. In < Imax in Figs. 7 and 8, however, In > Imax in Fig. 11. Assuming In > Imax, as shown in Fig. 11, the TEM operates in generation mode due to P < 0. At I = Ish, the TEM operates in short circuit cooling mode. At Ish < I < Imax and Imax < I < In, the TEM operates in efficient and inefficient cooling modes, respectively. When 0 < Th Tc < DTmax,c. The variation of current is at range from In to Isat. In this aspect, for In < I < Ii, a TEM operates in Inefficient cooling mode. Cooling mode exits for Ish < I < Ii because of Qc is positive bigger than Qc,open. Qc,open is cooling quantity under I = 0. Ii is defined as the current for which Qc equals its open circuit (I = 0 as Qc = Q c,open). So Ii = 2Imax. Certainly, ICOP=0.5 can be as a reference for evaluating the performance of TEC system. The optimal operating mode of TEM for cooling mode is at Ish < I < ICOP=0.5. It is defined that ICOP=0.5 is a threshold value. The major disadvantage of TEC system is that their thermodynamic efficiencies are smaller than those of competing technologies, such as vapor compression refrigerators. This is exacerbated by the fact that TEM normally require or produce low voltage/high current DC power because of the losses associated with voltage conversion. So it is beneficial for designers making operating mode of TEM towards a favorable direction. As shown in the Table 2, an efficient cooling operating mode could be acceptable in order to maintain the temperature of the object or taking away the cooling
ð24Þ
Table 2 Operating mode criteria and relationship of parameters in thermoelectric cooling system. Primary mode
Sub-mode
Qc
Temperaturedifference
P
I
Generation Neutral Cooling Cooling Neutral Cooling Heating Heating
+I generation Short circuit Efficient Inefficient – Inefficient – –
>0 >0 >0 >0 >0 >0 >0 >0 or < 0
Tc > T h Tc > Th Tc > Th Tc > Th Tc = Th Tc < Th Tc < Th Th > Tc or Th-Tc > DTmax
<0 =0 >0 >0 >0 >0 >0 >0
0 < I < Ish I = Ish Ish < I < Imax Imax < I < In I = In In < I < Ii Ii < I < Isat I > Isat Fig. 11. Variation of Qc, DT and COP with I at with extender block for Tj = 50 °C, E.
210
Y. Cai et al. / Energy Conversion and Management 124 (2016) 203–211
Fig. 12. Comparison between present analytical results and literature with different current.
load at required. Conditions parameters corresponding to minimize Tj or maximize Qc are considered though all conditions can’t reach at the same time.
Fig. 13. Comparison between present analytical results and measure with different current (E).
TEM module is considered from the parameters. It is seen that the Tj is obviously lower at L = 1 cm, which Rex is 0.03125 X. Also the analysis is agree with the measurement at 2–4 A. 4. Conclusions
3.3. Comparison with previous studies The present analysis method is compared with previous studies based on the two parts, one is for the thermoelectric properties, and the other is for chip temperature. Firstly, the performance is enhanced due to the effect of Seebeck coefficient with increasing of ZT has been presented in reference [29]. An assuming future predicted performance of TECs was discussed, which has the effect of increasing COP dramatically. For example, some detailed parameters for ZT = 2 on account of changing in Seebeck coefficient S can be found. The interesting temperature point of Tj = 50 °C at 170 W and Tj = 85 °C at 217 W are obtained from Fig. 5, greater than which corresponding result is 128 W and 169 W in the literature [29]. It is about 30% rise for improving cooling capacity. Another comparison based on the study of Zhang [13,14] is presented to verify the reliability of the present work. In their work, the expression of chip temperature was derived under the thermoelectric properties. The experimental study was conducted based on a processor test platform. In order to verify the correctness of the current research methods for operating mode of TEM system. Some parameters covering with Rjc,c and uU h Ah and Ta should be set to the same as the references. The comparison results for reference are shown in Fig. 12. It is found that the Tj under Qc = 100 W or 140 W is fitting with literature. This indicates the method of e NTU applied in electrical devices in this work is reliable and acceptable for predicting future TEC system. Fig. 12 also shows a comparison between measured Qc by Zhang with analytical results at Tj = 25 °C. Clearly, the present analysis is according with the experimental measure within a certain error range.
Based on the thermoelectric theory and the method of e NTU, a novel operating mode of thermoelectric module are established for electronic devices cooling system. The main feature of the present analytical method is that it makes full use of TEM parameters and studies the scale of extender block under two set analytical condition, that is, the chip temperature Tj at a fixed cooling capacity Qc and Qc at a fixed Tj, respectively. Theoretical expression of Tj, Qc and DT are derived in terms of present analytical method. Thermoelectric cooling system has been studied based on the thermoelectric module properties, and the optimal current for maximum cooling capacity and minimum chip temperature can be obtained under corresponding fixed condition, such as Rex. Results have demonstrated that thermoelectric properties make a great effect on the cooling capacity and chip temperature. The extender block shows better thermal performance in cooling capacity than chip temperature, so do COP and temperature difference. Future predict due to enhancing ZT only by changing Seebeck coefficient indicates Seebeck coefficient can improve cooling capacity compared with electrical resistance and thermal conductivity. The operating modes of thermoelectric module covering with control parameter are developed for getting a favorable cooling mode in this analytical method. The COP and temperature difference are as a reference for cooling mode forecasting. The cooling mode of TEM are within Ish < Ii. Finally, the validation of the present analysis is also conducted compared with previous studies and through the infrared thermal imager. Measuring results suggest good agreement is achieved between measurement and analysis.
3.4. Experiment validation Acknowledgements The experimental measurement was conducted by corresponding experiment table, which the heat source was replaced. A full scale experiment is setting up on the account of schematic diagram of thermoelectric cooler as shown in the Fig. 1. Two set measure under L = 1 cm and L = 5 cm are presented for comparison with temperature distribution to record Tj. Given the limited space available, TEC current and Qc are fixed to obtained temperature. TEC 12706 was selected as the test. Fig. 13 shows a surface temperature of heat source comparison of L = 1 cm and L = 5 cm at specified Qc = 45 W. As previous analysis, the operating cooling mode of
Authors would gratefully acknowledge the financial support of National Natural Science Foundation of China, (NSFC, Grant No. 51304233, Multi-physics diffusion of leaked natural gas and backward time inverse identification of leakage sources; NSFC, Grant No. 51208192, Instability theory and inverse convection design of air flow patterns in the large space), Hong Kong Scholar Program (XJ2013042), China Postdoctoral Science Foundation (2014M560593), Qingdao Postdoctoral Science Foundation, Fundamental Research Funding Program for National Key Universities in
Y. Cai et al. / Energy Conversion and Management 124 (2016) 203–211
China, Postgraduate Innovation Program (YCXJ2016073). In addition, Fu-Yun Zhao would like to acknowledge the financial support of the Thousand Youth Talents Plan from the Organization Department of CCP Central Committee (Wuhan University, China, Grant No. 208273259), Hunan Provincial Natural Science Foundation for Distinguished Young Scholars supported by Hunan Provincial Government (Grant No. 14JJ1002, Multiple fluid mechanisms of urban ventilation and its safety through source identification), Hubei Provincial Natural Science Foundation of Hubei Province (Grant No. 2015CFB261, Multiple macro flow states of urban built environment), Scientific Research Foundation for the Returned Overseas Chinese Scholars, State Education Ministry (Grant No. 230303, inverse identification of unperceived hazardous sources coupling with organized safe evacuation in the public buildings), National Key Basic Research Program of China (973 Program, Grant No. 2014CB239203), and Twelve-Five National Supportive Plan from Ministry of Science and Technology of China (Grant No. 2011BAJ03B07). References [1] Hanreich G, Nicolics J, Musiejovsky L. High resolution thermal simulation of electronic components. Microelectron Reliab 2000;40:2069–76. [2] Bar-Cohen A, Kraus AD, Davidson SF. Thermal frontiers in the design and packaging microelectronic equipment. Mech Eng 1983;105:53–9. [3] Zhang HY, Pinjala D, Teo PS. Thermal management of high power dissipation electronic packages: from air cooling to liquid cooling. In: IEEE, 5th electronics packaging technology conference (EPTC). p. 620–5. [4] Chein R, Huang G. Thermoelectric cooler application in electronic cooling. Appl Therm Eng 2004;24:2207–17. [5] Wang J, Zhao X-J, Cai Y-X, Zhang C, Bao W-W. Experimental study on the thermal management of high-power LED headlight cooling device integrated with thermoelectric cooler package. Energy Convers Manage 2015;101:532–40. [6] Xu D. Thermoelectric refrigeration and applied technology. Shanghai: Shanghai Jiaotong University Press; 1992. [7] Hodes M. On one-dimensional analysis of thermoelectric modules (TEMs). IEEE Trans Comp Packag Technol 2005;28:218–29. [8] Huang BJ, Chin CJ, Duang CL. A design method of thermoelectric cooler. Int J Refrig 2000;23(3):208–18. [9] Chen J, Yu J, Ma M. Theoretical study on an integrated two-stage cascaded thermoelectric module operating with dual power sources. Energy Convers Manage 2015;98:28–33.
211
[10] Lineykin S, Ben-Yaakov S. User-friendly and intuitive graphical approach to the design of thermoelectric cooling systems. Int J Refrig 2007;30(5):798–804. [11] Manikandan S, Kaushik SC. Energy and exergy analysis of an annular thermoelectric cooler. Energy Convers Manage 2015;106:804–14. [12] Simons RE, Ellsworth MJ, Chu RC. An assessment of module cooling enhancement with thermoelectric coolers. J Heat Transfer-Trans ASME 2005;127(1):76–84. [13] Zhang HY. A general approach in evaluating and optimizing thermoelectric coolers. Int J Refrig 2010;33:1187–96. [14] Zhang HY, Mui YC, Tarin M. Analysis of thermoelectric cooler performance for high power electronic packages. Appl Therm Eng 2010;30:561–8. [15] Liu Di, Zhao Fu-Yun, Yang Hong-Xing, Tang Guang-Fa. Thermoelectric mini cooler coupled with micro thermosiphon for CPU cooling system. Energy 2015;83:29–36. [16] Liu D, Zhao FY, Tang GF. Modeling and performance investigation of a closedtype thermoelectric clothes dryer. Dry Technol 2008;26:1208–16. [17] Bergman TL, Inreopera FP, lavine AS, Dewitt DP. Fundamentals of heat and mass transfer. 7th ed. John Wiley & Sons; 2011. [18] Rowe DM. Modules, systems and applications in thermoelectrics. CRC Press; 2011. [19] Crane DT, Bell LE. Design to maximize performance of a thermoelectric power generator with a dynamic thermal power source. J Energy Res Technol 2009;131:012401-1–1-8. [20] Hendricks T, Karri N. Micro- and nano-technology: a critical design key in advanced thermoelectric cooling systems. J Electron Mater 2009;38 (7):1257–67. [21] Zhou Y, Yu J. Design optimization of thermoelectric cooling systems for applications in electronic devices. Int J Refrig 2012;35(4):1139–44. [22] Meng JH, Wang XD, Zhang XX. Transient modeling and dynamic characteristics of thermoelectric cooler. Appl Energy 2013;108:340–8. [23] Zhao D, Tan G. A review of thermoelectric cooling: materials, modeling and applications. Appl Therm Eng 2014;66:15–24. [24] Seifert W, Ueltzen M, Müller E. One-dimensional modelling of thermoelectric cooling. Phys Status Solidi A 2002;194(1):277–90. [25] Hendricks T. Thermal system interactions in optimizing advanced thermoelectric energy recovery systems. J Energy Resour Technol 2006;129:223–31. [26] Luo Z. A simple method to estimate the physical characteristics of a thermoelectric cooler from vendor datasheets. Electron Cool 2008;14:22–7. [27] Lineykin S, Ben-Yaakov S. Analysis of thermoelectric coolers by a spicecompatible equivalent 0 circuit model. IEEE Power Electron Lett 2005;3:63–6. [28] Elsheikh MH, Shnawah DA, Sabri MFM, Said SBM, Hassan MH, Bashir MBA, et al. A review on thermoelectric renewable energy: principle parameters that affect their performance. Renew Sustain Energy Rev 2014;30:337–55. [29] Phelan PE, Chiriac VA, Lee TY. Current and future miniature refrigeration cooling technologies for high power microelectronics. IEEE Trans Comp Pack Technol 2002;25:356–65.