Professional Documents
Culture Documents
discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/51818219
CITATIONS READS
15 592
4 authors, including:
Debendra K. Das
University of Alaska Fairbanks
65 PUBLICATIONS 2,439 CITATIONS
SEE PROFILE
All content following this page was uploaded by Debendra K. Das on 18 March 2015.
In this study the electrical conductivity of aluminum oxide (Al2 O3 , silicon dioxide (SiO2 and zinc
oxide (ZnO) nanoparticles dispersed in propylene glycol and water mixture were measured in the
temperature range of 0 C to 90 C. The volumetric concentration of nanoparticles in these fluids
ranged from 0 to 10% for different nanofluids. The particle sizes considered were from 20 nm to
70 nm. The electrical conductivity measuring apparatus and the measurement procedure were val-
idated by measuring the electrical conductivity of a calibration fluid, whose properties are known
accurately. The measured electrical conductivity values agreed within ±1% with the published
data reported by the manufacturer. Following the validation, the electrical conductivities of different
nanofluids were measured. The measurements showed that electrical conductivity of nanofluids
increased with an increase in temperature and also with an increase in particle volumetric concen-
tration. For the same nanofluid at a fixed volumetric concentration, the electrical conductivity was
found to be higher for smaller particle sizes. From the experimental data, empirical models were
developed for three nanofluids to express the electrical conductivity as functions of temperature,
volumetric concentration and the size of the nanoparticles.
Keywords: Electrical Conductivity, Nanofluids, Nanoparticles, Particle Size, Propylene Glycol,
Volumetric Concentration.
RESEARCH ARTICLE
1. INTRODUCTION direct cooling. Nemchinsky and Severance5 describe that
an important factor influencing the erosion of cathode in
In the past decade there has been a substantial amount of plasma arc cutting is the cooling of the electrode. The bet-
research on nanofluids, which are dispersions of nanoscale ter it is cooled the lower is the erosion rate. On the higher
particles in regular fluids. Extensive experimental and the- power side, there are mechanized cutting designed to use
oretical studies including those of Jwo et al.,1 Yu et al.,2 high currents of the order of 400 A for oxygen cutting and
Vajjha et al.3 have established that nanofluids have sub- about 1000 A for argon–hydrogen mixture cutting. Water
stantially higher thermal conductivity than their baseflu- is commonly used as coolant but for operation at low tem-
ids, in which the particles are dispersed. An impact of peratures, ethylene glycol or propylene glycol and water
this phenomenon is reflected in enhancing the convective mixtures are necessary.
heat transfer coefficient of nanofluids. Therefore, nanoflu- The electrodes are hollow milled and a liquid tube is
ids can be used as a successful coolant in many indus- located in the center to provide a high flow velocity of
trial applications, such as the plasma arc cutters (PAC),
the coolant in the passage. The coolant convects away
which operate at very high temperature of the order of
the high degree of heat from the interior of the electrode
25000 K (Ramakrishna and Rogozinski).4 For many mate-
and torch parts. In this type of torch design, the nozzle
rials their thermal and electrical conductivities behave in
which is the outer annular structure surrounding the elec-
similar manners. Therefore, it is expected that just as the
trode is also cooled by liquid. As the demand on thickness,
thermal conductivity of nanofluid is higher than their base-
speed, strength and precision of metal cutting advances,
fluid, similarly the electrical conductivity of nanofluids
the heat flux generated within the metal body of PAC
may be higher than those of their basefluids.
grows in intensity. Therefore, better heat transfer fluids
In plasma arc cutters, the electrodes generate the plasma
are necessary to prevent possible overheating. An over-
arc and are subjected to intense heat and must require
heated plasma arc cutting system will wear out the torch
parts prematurely and may burn out the torch. Therefore,
∗
Author to whom correspondence should be addressed. nanofluids may be better coolants than their basefluid for
this application. They will be able to maintain the temper- summation of two conductivities due to ionic motion and
atures of the cathode and nozzle at the desired level at a particle motion and was expressed as
lower flow rate thus needing smaller passages and lower
3
pumping power. On the other hand, the metal cathode and nf = bf 1 − + Ze (1)
nozzle can be cooled to a lower temperature by nanofluids 4r 3
at the same volumetric flow as conventional coolant, thus where Z is the valence of a charged particle, e is the charge
diminishing the erosion rate. on an electron, is the electrophoretic mobility of charged
Ramakrishna and Rogozinski4 explain that for most effi- particles and r is the radius of particle. They made mea-
cient performance of plasma arc cutters, the electric power surements using -alumina powder of median particle size
(V × I) of the arc (V is the voltage of the plasma at the of 200 nm. They used water as the base fluid with addi-
nozzle exit with respect to cathode and I is the arc current) tions of 1N-HCl or 1N-NH4 OH to vary the pH of nanoflu-
should go into gas heating within the torch, without being ids between 2.0 to 10.0. Their measurements showed that
lost to the nozzle wall. Therefore, there is a possibility that the electrical conductivity was a function of the pH of
the higher electrical conductivity of nanofluids may affect the base fluid and nanofluid. The values of electrical con-
this electrical power. ductivity of nanofluid nf were higher than those of the
Since the fluid flows in the passages in metallic cathode basefluid bf , due to the contribution of charged mobile
and nozzle, which serves as the anode, the electrical con- alumina particles. Both values of nf and bf showed a
ductivity of the liquid may influence the life span of elec- minimum at pH in the range of 7 to 8. The electrophoretic
trodes. Therefore, it is essential that the designers know mobility and the valence of charged alumina particle Z
the electrical conductivity values of nanofluids accurately. varied with pH. There were no measurements presented on
With this knowledge, trade-off studies can be conducted by the effect of the temperature in their study, therefore, their
comparing the advantage gained by more efficient cooling measurements were possibly taken at room temperature.
of electrodes versus the effect of higher electrical conduc- Hayashi8 verified the electrical conductivity–
tivity on erosion of electrodes. temperature relation for natural waters proposed by
To fulfill this need, good data on electrical conductivity Sorensen and Glass,9 which was linear in a temperature
of nanofluids as a function of nanoparticle material, aver- range of 0 to 30 C.
age particle size, volumetric concentration and tempera-
ture are necessary. However, such information on electrical t = 25 1 + at − 25
(2)
RESEARCH ARTICLE
increase in the NH4 Cl additions. For the nanofluid with- Ganguly et al.13 measured the electrical conductivity of
out any NH4 Cl addition the conductivity increased with Al2 O3 nanoparticles in deionized water of average parti-
the increase in volumetric concentration of particles. By cle size 13 nm up to a volumetric concentration of 3%
classifying the particles under three categories, p bf , within the temperature range of 25–45 C. They expressed
p = bf , p bf they examined the interaction potential that the electrical conductivity of the nanofluid increased
between a pair of particles. They presented the conditions with an increase in temperature, and also it increased with
of ionic strength that would make the suspensions very sta- an increase in the volumetric concentration. They men-
ble, and the ionic conditions that would make the suspen- tioned that the Maxwell equation developed for electri-
sion move toward coagulation. Their results did not show cal conductivity of random suspension of microparticles
any measurements of electrical conductivity of suspen- underpredicted the values when applied to suspensions of
sions at different temperatures. Cruz et al.6 also presented nanoparticles. From their experimental data they presented
Maxwell’s model for the electrical conductivity of suspen- an empirical correlation valid for the Al2 O3 nanofluid in
sions for spherical particles, which is given as Eq. (4). the temperature range of 25 C to 45 C given by
nf 2bf + p − 2bf − p nf − bf
= (4) = 3679049 + 1085799T − 436384 (7)
bf 2bf + p + bf − p bf
This model is based upon the assumption that the parti- where T is in C.
cles are randomly distributed and at distances much larger
than their size. Therefore, there is no interaction among
them and the electrical field surrounding each particle is 3. EXPERIMENTAL SETUP AND
undisturbed. PROCEDURE
Abedian and Baker10 have presented a correlation for
3.1. Material Preparation
electrical conductivity of single phase liquid by measuring
the electrical conductivity and viscosity of five liquids used In this study three types of nanofluids, namely Al2 O3 , SiO2
as gas turbine lubricants. The equation they presented is and ZnO nanoparticles dispersed in PG/W (60:40) were
2ZCo used. Original concentrated aqueous suspensions of afore-
= (5) mentioned nanofluids were procured from Alfa Aesar14
R
and Nanostructured and Amorphous Materials, Inc.15 The
where viscosity = AeB/T −To
RESEARCH ARTICLE
(6) characteristics of materials used in our experiments are
In Eq. (5) Z, Co and R are ionic valence, concentra- tabulated below.
tion of charged species, a constant equal to 4 for poly- From these concentrated parent nanofluids, test sam-
mer solution and mean molecular size respectively. The ples of different volumetric concentration were prepared.
constants A and B in Eq. (6) are obtained from viscosity The exact mass of PG/W (60:40) mixture was calculated
measurements. to attain a desired volumetric concentration of nanofluid.
Glover et al.11 measured the electrical conductivity of Next, using a precise electronic mass balance, the calcu-
single wall carbon nanotubes dispersed in 50% water and lated amount of PG/W (60:40) mixture was added to the
50% ethylene glycol mixture. They observed a linear trend concentrated nanofluid by pipettes to reach the exact level
in the increase of the electrical conductivity with increase of particle volumetric concentration. In this way nanofluid
in weight percentage of carbon nanotubes. A 0.5 weight samples starting with 1% volumetric concentration to a
percent nanofluid exhibited 13 times the electrical conduc- maximum of 10% could be prepared. It was only the high
tivity of the base fluid. No measurements with temperature concentration of 45 nm Al2 O3 parent fluid with the ini-
variation were noted. tial concentration of 50% that yielded a 10% concentration
Wong and Kurma12 pointed out that presently there was sample. The others yielded sample concentrations of ≤6%.
insufficient data published on electrical conductivity of
nanofluids and good data were needed for an accurate 3.1.1. Sonication
determination of this property. To initiate research on this
property, they performed measurements of electrical con- The nanofluid samples were subjected to ultrasonication
ductivity of alumina nanofluid containing mean particle in a Branson Model 551017 sonicator under a frequency
diameter of 36 nm in deionized water as the base fluid. of 40 kHz. The ultrasonicator bath was filled with water
Their results showed the electrical conductivity of alu- up to the designated operating level. Then the water
mina nanofluid increased almost linearly with increase in was degassed for 5 minutes for removal of dissolved
the volumetric concentration of particles. No temperature gases as instructed by the manufacturer. Subsequently, the
dependency was studied by them and no correlation for nanofluid sample was placed in the water bath and son-
the electrical conductivity of the nanofluid was presented icated for about 3 hours to ensure uniform dispersion of
by them. nanoparticles.
ments were taken. In the present study, PG/W (60:40) was used as the base-
fluid to prepare different types of nanofluids. Therefore,
3.3. Benchmark Test Case the electrical conductivity of this base fluid was measured
first, which represented the conductivity values when the
The benchmark test of the electrical conductivity meter nanoparticle concentration is zero. The results of this mea-
and the probe was performed using the conductivity cal- surement are shown in Figure 3. A second order polyno-
ibration liquid of Hanna designated as HI 7033. These mial relation fits the data well.
electrical conductivities measured at various temperatures
of the calibration liquid have been plotted in Figure 2. bf = 2316 × 10−5 T 2 − 1066178 × 10−2 T
For comparison the conductivity values of this liquid pro-
+ 127618050 with R2 = 09998 (7)
vided by the manufacturer have also been plotted in this
figure. It was observed that the measured values of the
electrical conductivities and the values published by the 0.5
Electrical conductivity σbf μS\cm
PG/W(60:40)
manufacturer differed by about ±1%. Thus the benchmark 0.45
Polynomial fit
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
270 280 290 300 310 320 330 340 350 360 370
Temperature K
Fig. 1. Experimental setup for measuring the electrical conductivity of Fig. 3. Electrical conductivity of PG/W (60:40) as a function of
nanofluids. temperature.
This equation can be nondimensionalized by the electri- nanofluid with 20 nm particle size increases by 106%,
cal conductivity bfo of the basefluid at a reference temper- when the volumetric concentration increases over the same
ature To , which was adopted to be the room temperature range from 1 to 4%. For about 30% increase in tem-
of 20 C. perature, the nf increases by 400%, but for a 300%
2 increase in , nf increases by 100–200%. Therefore, the
bf T T
= 1414710791 − 2222753013 variation of electrical conductivity of nanofluid shows a
bf0 T0 T0 stronger dependence on temperature than the volumetric
+ 9080421369 with R2 = 09998 (8) concentration within the ranges of our experiments. It is
observed that for the same particle volumetric concentra-
The Eq. (8) is valid in the range 273 K ≤ T ≤ 363 K. tion, at a fixed temperature, nanofluids containing smaller
nanoparticles exhibit higher electrical conductivity than
4.2. Aluminum Oxide Nanofluid those containing larger nanoparticles. This is due to the
presence of more number of charged particles of smaller
Figure 4 presents a comparison between the electrical con- size for the same volumetric concentration. Comparing
ductivity of the Al2 O3 nanofluid in PG/W (60:40) of three the trends of measured data in Figure 4 with the equa-
average particle sizes, namely, 10, 20 and 45 nm. Sev- tion Wang and Hirata,7 Eq. (1), we observe an agreement.
eral distinct characteristics are observed. The electrical Their equation predicts that the electrical conductivity of
conductivity of Al2 O3 nanofluids increases as the temper- a nanofluid increases with an increase in concentration
ature increases. The electrical conductivity shows a lin- and a decrease in particle size. Our data exhibit the same
ear variation with temperature. As an example for the trend.
20 nm nanofluid of 1% concentration, the linear varia-
tion of electrical conductivity with the temperature is given
by Eq. (9). 4.3. Silicon Dioxide Nanofluid
nf = 13732T − 35539 with R2 = 09923 (9) Figure 5 shows the variation in electrical conductivity of
the SiO2 nanofluid with temperature. Only one average
The Eq. (9) is valid in the range 273 K ≤ T ≤ 363 K. particle size of 30 nm for this nanofluid was available from
The electrical conductivities of 1% Al2 O3 nanofluid the manufacturer. Similar to the observation for Al2 O3
of particle sizes of 10, 20 and 45 nm increased by nanofluids, the electrical conductivity of SiO2 nanofluid
RESEARCH ARTICLE
441%, 444% and 416% respectively, when the tempera- also increases with an increase in temperature in a lin-
ture increased from 273 K to 363 K (34%). At a specific ear fashion for a constant concentration, in the tempera-
temperature, as the volumetric concentration of nanofluid ture range of 273 to 363 K. As an example, the electrical
increases, its electrical conductivity increases. For exam- conductivity data of the 3% SiO2 nanofluid matches the
ple at a temperature of 50 C, the electrical conduc- expression
tivity of the Al2 O3 nanofluid with an average particle
size of 45 nm increases by 191% when the volumet- nf = 25241T − 64104 with R2 = 099 (10)
ric concentration increases from 1% to 4%. At the same
temperature of 50 C, the electrical conductivity of this valid in the range 273 K ≤ T ≤ 363 K.
700 450
1% Al2O3 10 nm 1% Al2O3 20 nm 1% SiO2 30 nm
Electrical conductivity σnf μS\cm
300 200
150
200
100
100
50
0 0
270 290 310 330 350 370 270 280 290 300 310 320 330 340 350 360 370
Temperature K Temperature K
Fig. 4. Comparison between the electrical conductivity of the Al2 O3 Fig. 5. Electrical conductivity variation with temperature for the SiO2
nanofluid as a function of temperature containing 10, 20 and 45 nm nanofluid in PG/W (60:40) basefluid for particle concentration ranging
particles. from 0 to 6%.
140 3% ZnO 70 nm
4% ZnO 70 nm Table I. Material characteristics of nanofluids used in the present
120
1% ZnO 36 nm experiments.
2% ZnO 36 nm
3% ZnO 36 nm
100 4% ZnO 36 nm Parent Particle
0% ZnO (PG/W) Particle Particle nanofluid electrical
2nd order poly. fit
80 size density concentration conductivity16
Manufacturer Material nm g/cc wt% in H2 O S/cm at 20 C
60
Alfa Aesar Al2 O3 20 3.614 30 102 × 10−10
40 Alfa Aesar Al2 O3 45 3.6 50 102 × 10−10
Nanostructured
−Al2 O3 10 3.615 20 102 × 10−10
20 and Amorphous
Materials, Inc.
0
270 280 290 300 310 320 330 340 350 360 370 Nanostructured SiO2 30 2.41515 25 10−9
and Amorphous
Temperature K
Materials, Inc.
Alfa Aesar ZnO 36 5.614 40 —
Fig. 6. Comparison between the electrical conductivity of ZnO
Alfa Aesar ZnO 70 5.6 50 —
nanofluid as a function of temperature of particle sizes 36 and 70 nm.
Table II. Regression coefficients of the electrical conductivity correla- conductivity ratio with the particle size, the experimental
tion (Eq. (16)) for different nanofluids.
values of nf /bf of each nanofluid were plotted against
Regression constants ZnO SiO2 Al2 O3 its nondimensional average particle sizesd0 /d. The par-
ticle size d varied from 10 to 70 nm and d0 , the reference
a1 −8177.324 2928.485 −1772.883
a2 1413.054 23095.615 1128.208 average particle size was adopted to be 100 nm, which
a3 2.2848 419.136 14.425 is generally accepted as the upper limit for the nanoscale
b1 −2.719 −3.373 −2.069 range. From these plots it was determined that variation of
b2 5.594 7.3092 4.578 the electrical conductivity with an average particle size for
b3 −2.584 −3.3397 −2.204
each nanofluid followed a linear relation
c1 11.681 0a 11.456
c2 8.383 1a −16.256 nf d
R2 0.98 0.98 0.98 = c1 0 + c2 (15)
Maximum deviation% −15% +10% +20% bf d
Average deviation% −1.74% 3.12% 3.93%
From these individual analyses of the influence of each
a
Since only one particle size was tested for SiO2 nanofluid the diameter effect is parameter, it was concluded that a correlation in the fol-
not included.
lowing format would be appropriate for the electrical
those parameters independently. To make the correlation conductivity.
independent of units, the electrical conductivity nf was 2
nf T T
nondimensionalized by bf using Eq. (8). = a1 + a2 + a3
b1
2
+ b2 + b3
bf T0 T0
4.6.2. Influence of Concentration d
× c1 0 + c2 (16)
d
From Figure 7 and equation Eq. (12) it was observed that
the electrical conductivity nf variation as a function of Next the statistical package LAB fit21 was used to deter-
followed a polynomial function of second degree at a mine the unknown regression coefficients of Eq. (16)
constant temperature and particle size. This was tested for for different nanofluids from their experimental data. The
each of the nanofluids at several fixed temperatures and regression coefficients from the statistical analysis are tab-
particle sizes of 10, 20, 30, 36, 45 and 70 nm. It was found ulated in Table II.
that the nature of the correlation can be represented the The Eq. (16) with coefficents of Table II has the range
RESEARCH ARTICLE
best way by of validity;
nf
= a 1 2 + a2 + a3 (13) 273 K ≤ T ≤ 363 K; 1% ≤ ≤ 6%; 10 nm ≤ d ≤ 70 nm.
bf
4.7. Comparison of Experimental Values with
4.6.3. Influence of Temperature Correlation
The electrical conductivity nf data of each of nanofluid Comparison of the electrical conductivity values predicted
in Figures 4–6 were nondimensionalized with the base- by the correlation (Eq. (16)) and the experimental val-
fluid electrical conductivity bf and were plotted against ues for the ZnO nanofluid is shown as an example in
the nondimensionalized temperature T /T0 , for fixed con- Figure 8 using the software Matlab.22 The central line
centration and fixed particle sizes. It was observed that the
electrical conductivity ratio nf /bf followed a second
order polynomial in T /T0 . A careful analysis into this
variation confirmed that although nf was a linear function
of T , bf was nonlinear in T as observed in Figure 3. Due
to a higher rate of increase of bf with T , the best fit cor-
relation for nf /bf as a function of T /T0 emerged to
be of the form
2
nf T T
= b1 + b2 + b3 (14)
bf T0 T0
represents a perfect match between the experimental and p electrical conductivity of particle material, S/cm
correlation values and the two dashed lines are the 95% volume fraction of nano particles
prediction bound. Only 2 data points out of about 100 data
points fall out side the 95% prediction bounds on the
upper side. Given a specific concentration, diameter and References and Notes
temperature of this nanofluid within the curve-fit range, 1. C. S. Jwo, H. Chang, T. P. Teng, M. J. Kao, and Y. T. Guo,
the correlation ensures 95% confidence that the electrical J. Nanosci. Nanotechnol. 7, 2161 (2007).
conductivity value will be between the lower and upper 2. W. Yu, D. M. France, D. Singh, E. V. Timofeeva, D. S. Smith, and
prediction limits. J. L. Routbort, J. Nanosci. Nanotechnol. 10, 4824 (2010).
3. R. S. Vajjha and D. K. Das, Int. J. Heat Mass Transfer 52, 4675
(2009).
5. CONCLUSIONS 4. S. Ramakrishnan and M. W. Rogozinski, J. Phys. D: Appl. Phys.
30, 636 (1997).
The electrical conductivity of three nanofluids were 5. V. A. Nemchinsky and W. S. Severance, J. Phys. D: Appl. Phys.
39, R423 (2006).
measured and found to be functions of volumetric
6. R. C. D. Cruz, J. Reinshagen, R. Oberacker, A. M. Segadaes, and
concentration, temperature and particle diameter. The elec- M. J. Hoffmann, J. Colloid Interface Sci. 286, 579 (2005).
trical conductivity exhibited a strong dependence on tem- 7. X. H. Wang and Y. Hirata, Journal of Ceramic Processing Research
perature. For the Al2 O3 nanofluid an increase of 400% in 1, 64 (2000).
electrical conductivity was observed for a temperature rise 8. M. Hayashi, Environ. Monit. Assess. 96, 119 (2004).
of about 30%. For the same nanofluid, an increase of about 9. J. A. Sorensen and G. E. Glass, Analyt. Chem. 59, 1594 (1987).
10. B. Abedian and K. N. Baker, IEEE 3, 888 (2008).
100–200% in electrical conductivity was observed for a 11. B. Glover, K. W. Whites, H. Hong, A. Mukherjee, and W. E. Billups,
volumetric concentration increase of 300%. The electrical Synthetic Materials 158, 506 (2008).
conductivity was found to be inversely proportioned to par- 12. K. F. V. Wong and T. Kurma, Nanotechnology 19, 345702
ticle diameter. From experimental data a correlation was (2008).
developed (Eq. (16)) that is suitable for the prediction of 13. S. Ganguly, S. Sikdar, and S. Basu, Powder Technology 196, 326
(2009).
the electrical conductivity of these nanofluids. With future 14. Alfa Aesar. Available from: <http://www.alfaaesar.com/>, last
research this correlation can be improved to be valid for accessed 2007.
other nanofluids. 15. Nanostructured and Amorphous Materials, Inc. Available from:
<http://www.nanoamor.com/>, last accessed 2010.
RESEARCH ARTICLE
Nomenclature 16. D. R. Lide, Handbook of Chemistry and Physics, 84th edn., CRC
Press, Boca Raton, FL (2003).
d average particle size, nm 17. Bransonic Tabletop Ultrasonic Cleaners, Branson Ultrasonic Corpo-
ration, Danbury, CT (2010).
do reference average particle size, 100 nm
18. Experimental Operating and Maintenance Procedures for Electrical
T absolute temperature, K Conductivity of Liquids Unit, Hanna Instruments, Woonsocket, RI
To reference temperature = room temperature, (2009).
20 C (293 K) 19. VWR Model 320 Heater, Available from: <http://www.vwr.com/>,
last accessed 2009.
Greek symbols 20. Environmental Test Chamber Model FS 202 chamber 144, Associ-
ated environmental systems, Lawrence, MA (2002).
21. W. P. Silva and C. M. D. P. S. Silva, LAB Fit Curve Fitting
nf electrical conductivity of nanofluid, S/cm
Software (Nonlinear Regression and Treatment of Data Program)
bf electrical conductivity of basefluid, S/cm V 7.2.42 (1999–2008), online, available from world wide web:
bfo electrical conductivity of basefluid at reference <www.labfit.net>, date of access: April (2008).
temperature, S/cm 22. Matlab R2008a, MathWorks, Natick, MA (2008).