Thermal Conductivity of Ionic Liquids and ... - ACS Publications

past eight years, work has been developed by Nieto de Cas- tro et al.8-14, by successfully measuring density, .... Multi-walled carbon nanotubes (MWCN...
1 downloads 9 Views 2MB Size
Subscriber access provided by UNIV OF DURHAM

Thermodynamics, Transport, and Fluid Mechanics

Thermal Conductivity of Ionic Liquids and IoNanofluids and their Feasibility as Heat Transfer Fluids João Manuel Pedro Moisão França, Maria José V. Lourenço, S.M. Sohel Murshed, Agilio A. H. Padua, and Carlos A. Nieto de Castro Ind. Eng. Chem. Res., Just Accepted Manuscript • Publication Date (Web): 16 Apr 2018 Downloaded from http://pubs.acs.org on April 16, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

Thermal Conductivity of Ionic Liquids and IoNanofluids and their Feasibility as Heat Transfer Fluids João M.P. França†,‡,# , Maria José V. Lourenço†, S.M. Sohel Murshed†, Agílio A.H. Pádua‡, Carlos A. Nieto de Castro †,* † ‡

Centro de Química Estrutural, Faculdade de Ciências, Universidade de Lisboa, Campo Grande, 1749-016, Lisboa, Portugal. Institut de Chimie de Clermont-Ferrand, Université Clermont Auvergne & CNRS, 63178 Aubière, France.

KEYWORDS: ionic liquids, IoNanofluids, thermal conductivity, heat transfer area. ABSTRACT: Ionic liquids and IoNanofluids were studied in recent years as possible alternatives to current engineering fluids, namely in the area of heat transfer. Excellent thermal properties, like high heat capacity per unit volume and thermal conductivity, allied to the dispersion of nanoparticles in them, have created great expectations, as the enhancement of their thermophysical properties liquids can contribute to better efficiency in heat transfer. The thermal conductivity of [P66614][N(CN)2], [P66614][Br], [C2mim][SCN], [C4mim][SCN], [C2mim][C(CN)3] and [C4mim][C(CN)3] in the temperature range 293 to 343 K at 0.1 MPa and their IoNanofluids with MWCNTs is reported in the present work. While we could not obtain stable suspensions with phosphonium based ionic liquids, thermal conductivity enhancement of cyano-based ionic liquids was compared with our previous work using dicyanamide ionic liquids. The thermal conductivity of C2mim+ ionic liquids and IoNanofluids is generally higher than the corresponding C4mim+ fluids. Temperature dependence of thermal conductivity enhancement hinders the conception of a unified thermal conductivity enhancement predictive model of the presented IoNanofluids, current theories under-predicting its value for the dispersions studied. Finally, we selected a specific heat transfer process and calculated the heat transfer area necessary using currently commercialized heat transfer fluids, ionic liquids and IoNanofluids. While the addition of nanomaterial to the ionic liquids leads to an increase in the heat transfer available area, the enhancement of the thermophysical properties leads to a smaller variation of the area with temperature. Depending on the ionic liquid, some of the IoNanofluids studied are head-to-head with a significant number of currently used heat transfer fluids concerning the heat transfer area necessary to transfer the same amount of heat.

INTRODUCTION

Environmental awareness and fluid process energy optimization made way for many applications in modern chemistry and process engineering. The unique characteristics of ionic liquids allow us to consider them beneficial for the environment, since they can be used as green solvents.1,2 These low temperature liquid salts have drawn the attention of scientific and industrial communities as possible new heat transfer fluids (HTFs), possible replacements of environmentally harmful ones in use by industry, with maximum possible energy efficiency, for healthy and sustainable industrial and domestic applications. In addition, the distinct nanomaterials properties (mechanical, optical, magnetic, electrical, chemical and thermal)3 can enhance the thermal properties of stable dispersions in base fluids, called nanofluids4, and once studied their toxicological effects5,6, small quantities of them can greatly enhance the properties of the base fluid. The amount of nanomaterial used in our work (nanocarbons) is small and was handled without contact to skin or inhalation.

If the base fluid is an ionic liquid, the use of well characterized nanoparticles permits the design of flexible soft materials having the required properties for a certain duty. Considering applications in heat transfer, ionic liquids have fundamental properties high heat capacity per unit volume and thermal conductivity,7 which permit its consideration as possible HTFs. The dispersion of nanomaterials in these fluids created a great expectation, as their thermophysical properties are increased in relation to the base fluid. Over the past eight years, work has been developed by Nieto de Castro et al.8-14, by successfully measuring density, heat capacity, thermal conductivity and viscosity of ionic liquids containing imidazolium with suspended multi-walled carbon nanotubes (MWCNTs), as a function of temperature. The term IoNanofluids (INFs) was proposed, to designate a kinetically stable suspension of nanoparticles in ionic liquids.8 The preparation of nanofluids has been discussed in a previous publication of the Lisbon group.15 It is a fundamental step concerning stability, characterization and use. A recent analysis of the possible steps for nanofluid preparation and characterization, including recommended stability tests, showed that it is possible to prepare stable nanofluids (over a period of several months to years)16. The viscosity of these IoNanofluids has to be reasonably small, and in our latest

1 ACS Paragon Plus Environment

Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

work on this topic,13 we discussed the use of dicyanamide (alkylimidazolium and alkylpyrrolidinium) based IoNanofluids in heat transfer since ionic liquids based on the dicyanamide anion have lower viscosities than most of the known ionic liquids. In this paper data on cyano alkylimidazolium ionic liquids, such as the ones based on thiocyanate (SCN-) and tricyanomethanide (C(CN)3-) anions, that have similar viscosity values,17 is reported. Additionally, these anions tend to have a lower ecotoxicity than anions such as (CF3SO2)2N- and tris(perfluoroalkyl)trifluorophosphate),18-19 which is an advantage from an environmental point of view. Unfortunately, the same cannot be said regarding the alkylimidazolium cations. However, the selection of smaller side chains reduces the biological effects and, in this sense, an acceptable cost-benefit ratio can be attained from the use of these fluids. Besides viscosity, the fluid heat capacity will have a significant effect on its viability as a heat transfer fluid. As it has been mentioned in previous work,7 heat transfer area is the main parameter used for the heat exchangers design, estimated from the thermophysical properties of the selected HTF. The inner heat transfer coefficient (hi) of a shell and tube heat exchanger, which conditions the heat transfer area, is directly proportional to the thermophysical properties of a fluid:



(1)

Although the heat capacity has the smallest influence in this relation, a viable heat transfer fluid could be obtained if the latter were to be great enough to account for any variations of the remaining properties. Such is the case for phosphonium cation based ionic liquids, with greater CP values when compared with their alkylimidazolium counterparts.20-22 Even though the viscosity values are much greater in the case of ionic liquids with [P66614] cation, we found it to be valuable to explore these type of ionic liquids and their IoNanofluids as potential HTFs. This work reports measurements of the thermal conductivity of trihexyltetradecylphosphonium dicyanamide ([P66614][N(CN)2]), trihexyltetradecylphosphonium bromide ([P66614][Br]), 1-ethyl-3-methyl-imidazolium thiocyanate ([C2mim][SCN]), 1-n-butyl-3-methyl-imidazolium thiocyanate ([C4mim][SCN]), 1-ethyl-3-methyl-imidazolium tricyanomethanide ([C2mim][C(CN)3]) and 1-n-butyl-3-methylimidazolium tricyanomethanide ([C4mim][C(CN)3]) in the temperature range 293 to 343 K, at 0.1 MPa, and their IoNanofluids with MWCNTs. Results obtained were compared with those predicted by current theories of thermal enhancement in nanofluids. Finally, we evaluated the heat transfer area necessary for a given process using currently commercialized heat transfer fluids (HTFs), ionic liquids or IoNanofluids and ponder the pros and cons of using these substances as new heat transfer fluids.

Page 2 of 28

MATERIALS AND METHODS Materials Ionic liquids were obtained from IoLiTec, Ionic Liquids Technologies, Germany, specifications from certificate analysis given in Table S1 of the Supporting Information. Their estimated purity (mass fraction) was 0.98 or better. The presence (below 2%) of the bromide precursor used in their synthesis was responsible for a pale light color shown by them. Several minor impurities at low concentrations, such as halides, will not play a significant role on the thermal conductivity values.23 However, the ionic liquids water content is relevant since it may change considerably the properties of the liquid, in particular for the thermal conductivity12 when above 1000 ppm. Further purification was performed by drying under vacuum the ionic liquids at T = 333K, for a period not less than 72 h. Given our experience with these substances and methods from our previous work, we can confidently assume the water content was minimized to a point where its presence did not affect the presented results. Another factor favoring of this statement lies in the extra 24 h of drying under vacuum instead of the typically used 48 h. Multi-walled carbon nanotubes (MWCNTs), a development product previously used by our group,8 were obtained from Bayer Material Science (Baytubes C150 HP). The characteristics of this material are displayed in Table S2 in the Supporting Information.

Equipment and methodology

Thermal conductivity measurements were obtained with the KD2 Pro™ Thermal Properties Analyzer (Decagon Devices, Inc.). The principle of measurement, based on the transient hot-wire method (THW), uses a single-needle sensor (KS-1, 1.3 mm diameter and 60 mm long), to be inserted vertically in the sample to minimize free convection, contains a heating element and a thermal resistor. The measurement is accomplished by heating the sensor at a certain rate, monitoring its temperature variation as a function of time with the thermal resistor simultaneously. Data acquisition is controlled by a microprocessor connected to the sensor. The sample thermal conductivity is calculated using a parameter-based corrected version of the temperature model given by Carslaw and Jaeger for an infinite line heat source with constant heat dissipation, with zero mass, immersed in an infinite medium,24 the transient hot-wire idealization. Further information about these calculations are available elsewhere.8, 25 To ensure that sensor dimensional requirements were realized in pratice, samples of the liquids with approximate volume of 85 cm3 were used in a suspended glass vial, to avoid transmission of vibrations of the temperature-controlled oil bath (Haake C25; oil Galp Electric 2). The sensor was inserted vertically, and temperature was allowed to stabilize prior to a sample measurement. Time to reach the desired temperature was sample dependent, longer for IoNanofluids (more viscous, smaller thermal diffusivity). 8 to 10 measurements were taken to ensure measurement reproducibility, with an

2 ACS Paragon Plus Environment

Page 3 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

interval of 15 min between them. Temperature was measured by the thermal conductivity sensor, with manufacturers quoted uncertainty u(T) = 0.1 K. Calibration of the device is very fundamental to guarantee its capabilities. The calibration procedure for the thermal conductivity measurements used here was extensively described elsewhere.12 Since our last publication on this matter, the software of the KD2 Pro™ was updated and a new calibration was performed exactly in the same way as before. For the calibration reference fluids (water, toluene, glycerin, a mixture of glycerin and water (50/50 w/w) and an aqueous solution of NaCl) were used to calculate the calibration constant, K, from the ratio between the experimental thermal conductivity (λmeas) and the selected reference data (λref) at the same temperature and pressure, using equation 2: (2) This constant, was found to be liquid and temperatureindependent, with a value of (1.028 ± 0.012), an overestimate (+2.8%) relatively to the reference values, requiring the necessary correction of the measured data. Reference values for the thermal conductivity of the calibrating fluids can be found in reference13. The measurement uncertainty depends on the fluids studied, namely from their viscosity. From our experience, the instrument used, due to its software conception, is better suited (lower uncertainty) for liquids with viscosities greater than 20-30 mPa.s. Lower viscosities increase the probability of heat transfer by convection, which has two immediate effects – the need for lower time in heating the transient hot wire probe, and greater time interval between consecutive measurements. Only this last one we could control, carefully screening very well the results obtained, and increasing the uncertainty in the measurements, as explaining in the results section.

Preparation of IoNanofluids

Based on previous work, IoNanofluids were prepared by adding a specific mass of nanotubes to the ionic liquids, followed by the dispersion with a sonication probe (Sonicator Sonics & Materials, VC50), with 0.5 and 1% mass fraction of MWCNTs. The sonication time used for each mass fraction was around 1h30, divided into several periods to allow the cooling of the sample. Following previous work,12-13 our goal was to obtain a possible HTF with some fluidity, not a gel.. Uncertainty in the IoNanofluids mass fraction was calculated to be u(w) = 0.0004. This preparation did not involve addition of surfactants to the dispersion, to avoid masking of the thermal conductivity values. In fact, Murshed et al.26 demonstrated for nanofluids based in water, that all surfactant-added nanofluids showed larger enhancements in thermal conductivity than the surfactant-less, the viscosity also becoming larger. In addition, nanoparticle clustering showed to have negative impact on the stability and thermal conductivity of these nanosystems. Demonstration that undesirable aggregates are not present in the prepared IoNanofluids can be performed with dynamical

light scattering (DLS) techniques, and also by TEM imaging and subsequent counting of number of particles with average sizes. Both these methods are important to ensure that nanomaterials and nanofluids are well understood, but have many limitations for nanofluids, namely for those that are non-light transparent like the carbon nanofluids in DLS. On the other hand, TEM imaging can characterise solids and even solidliquid interfaces, but never the liquid itself.27 Our previous work with other IoNanofluids suggest that aggregation at a microscopic scale is not present, as no sedimentation was observed for several months in the case of the SCN-, N(CN)2)- and C(CN)3- based IoNanofluids. The same could not be said for the phosphonium based IoNanofluids, as we will discuss below.

Theoretical Interpretation

Depending on nanoparticles (concentration, type, and morphology) and liquid type, the thermal conductivity enhancements of nanofluids relative to their base liquids varies considerably from moderate to large percentages.28-30 In addition there are other mechanisms, which also contribute to the enhancement of the thermal conductivity of IoNanofluids. Therefore and like nanofluids, the thermal conductivity of these IoNanofluids cannot be predicted by classical effective medium theory-based models, such as those attributed to Maxwell31 and Hamilton and Crosser.32 Hamilton and Crosser model is a modification of Maxwell’s model (applicable for spherical particles only), by applying a particle shape factor for suspensions of non-spherical particles. These classical models could not predict and explain the experimental thermal conductivity.28-29, 33 Despite a sizeable research effort devoted in the last decade to identify the mechanisms and to develop theoretical models for nanofluids,29-30, 34 our current knowledge on the real mechanism for heat transfer in nanofluids is scarce and sometimes controversial.28, 35 Nonetheless, there are a few main mechanisms that explain, at least qualitatively, the enhanced thermal conductivity of nanofluids. These include Brownian motion of nanoparticles (for spherical shape), interfacial nanolayer at the surface of nanoparticle/base fluid interface, and nanoparticle clustering or aggregation. Regarding the interfacial nanolayer, the liquid molecules near the particle surface interact and/or are absorbed at the nanoparticle surface creating a layered structure, which shows characteristics of an organized “solid” like system.34, 36-37 These mechanisms are well discussed in the literature29, 34, 36 and are not elaborated further here. Among these mechanisms, the interfacial nanolayer was found to the major factor behind such increase in thermal conductivity of nanofluids.34, 37-38 Unlike nanofluids, no attempt has so far been made on the theoretical development for the thermal conductivity of IoNanofluids prediction. Since the latter are a new class of nanofluids, these mechanisms are also presumed to be applicable for their thermal conductivity. The Hamilton and Crosser (HC) model32 for the thermal conductivity of suspensions (λeff) is a function of the thermal conductivities of both the particles (λp) and liquid (λf), vol-

3 ACS Paragon Plus Environment

Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ume fraction of particles (ϕp), and particle shape. The model is expressed as: ! "#

!

! "#

! "#$ % $ %

!

!

'

&'

(

)

$ *%

! * '+,-. ! , . "/ $ *% ! * '+,-. ! , . ,-. % * ' ! % ! * '$ +,-. , "/

"/

conditions), the heat transfer area can be calculated from general principles,41-42: 1

(3)

The shape factor is assumed to be n = 3 for spherical particles and n = 6 for cylindrical particles (e.g., nanotubes). Among a handful of efforts, Murshed et al. 38 developed two models accounting for the effects of particle size, concentration, and interfacial nanolayer for nanofluids (λeff-nf), for nanofluids with spherical and cylindrical nanoparticles. Their model (MLY) for nanofluids containing cylindrical/tube nanoparticles has the form: "



0#

In this equation, ω=λlr/λf, γ=1+h/rp, γ1=1+h/(2rp), rp is the nanoparticle radius, h and λlr are, respectively, the thickness and the thermal conductivity of interfacial nanolayer. Although the thickness of nanolayer can be considered to be 2 nm,30, 38 recent molecular dynamics studies showed that the interfacial layer is around 0.9 nm for cyano based IoNanofluids.39 The order and orientation of fluid molecules adsorbed on the nanoparticle surface result in an intermediate value (between solid and base fluid) for the thermal conductivity of nanolayer i.e., λf < λlr < λp. Therefore, the thermal conductivity nanolayer is given by λlr = ωλf, where ω >1 is an empirical parameter that depends on the fluid molecules organization in the interface, as well as on the nature and surface chemistry of nanoparticle. According to MLY model, the nanolayer thermal conductivity is several times the thermal conductivity of base liquid.

Heat transfer area calculation

The methods used to calculate the heat transfer area have been extensively described in previous publications,7, 40 where the changes in design parameters of shell and tube heat exchanger are only dependent on the variation of the HTFs thermophysical properties, namely density, heat capacity, thermal conductivity and viscosity. Heat exchanger degrees of freedom can be reduced to two if a specific type of device is selected and by specifying the required duty imposed by the external constraints of a particular process: heat transfer area and fluid ducts pressure drop. Since the latter is not usually considered a major factor in the design, the heat transfer area has been considered, as before, the sole factor reflecting the variation of the thermophysical properties of heat transfer fluid considered. The equipment chosen, described in references 7 and 40, is a solar power unit with a molten salt receiver as its thermal energy storage system. It captures solar energy, storing it in molten salts (sodium nitrate or molten nitrate mixtures), at high temperature, in order to generate power, independently of the daytime period. Using a set of reference conditions (and the thermophysical properties of each fluid at these

Page 4 of 28

2 3 45#6

(5)

In this equation Q is heat transfer rate, U0 the overall heat transfer coefficient, A0 the heat transfer area, (∆T)lm being the logarithm mean temperature difference between the two fluids inlet and outlet stream temperatures. Therefore, and for a given duty, the only design parameter that depends on thermophysical properties is the heat transfer area (for a shell and tube heat exchanger, the number of tubes and passages of fluid). The “barrier resistance” to heat transfer is measured by the overall heat transfer coefficient, which includes all the resistances to the latter, namely the contribution due to convection at the tubes’ inner and outer surfaces and conduction across the tube walls. Convective heat transfer is controlled by the surfaces boundary layers, which depend on the above mentioned fluid thermophysical properties. The overall heat transfer coefficient for circular tubes, can be calculated from: " 2

"

"7 7

89

8

8

7 7

(6)

D0 and Di are, respectively, the outside tube diameter and the inside tube diameter, rw is the tube wall thermal resistance, ri and r0 are the internal and external fouling resistances, and hi and h0 are the the inside and outside film fluids heat transfer coefficients. For the current application, equation (6) can be simplified to: " 2

"

"7 7

:

(7)

R represents the overall resistance of the tube wall, other factors being assumed constant. Optimal heat transfer is obtained with turbulent flow, which in normally achieved by using baffles in the shell part and by controlling fluid velocity in the tubes. Sieder and Tate correlation43 for a turbulent flow in a smooth circular tube is given by: ;
?:@

A8 -⁄ C D

-

(8)

Re and Pr are, respectively, the Reynolds and the Prandtl numbers (given by equations (9) and (10)), λ and η, the fluid thermal conductivity and viscosity evaluated at the fluid mean bulk temperature. CP is the heat capacity, ρ is the fluid density and u is the tube cross section mean fluid velocity. ηs is the fluid viscosity at wall temperature. The left-side of equation (8) is the Nusselt number. A8 :@

(9)