Study of a thermoelectric generator based on a catalytic premixedmeso-scale combustor

L. Merotto a,⇑, C. Fanciulli b, R. Dondè a, S. De Iuliis a

aCNR – IENI (Istituto per l’Energetica e le Interfasi), via Cozzi 53, 20125 Milano, ItalybCNR-IENI Corso Promessi Sposi 29, 23900 Lecco, Italy

h i g h l i g h t s

! A catalytic meso-scale combustor fueled with propane/air is developed and tested.! The combustor is coupled with thermoelectric modules for electric power production.! The system produces 9.86 W electrical power with an overall efficiency up to 2.85%.

a r t i c l e i n f o

Article history:Received 23 July 2015Received in revised form 17 September 2015Accepted 11 October 2015Available online 11 November 2015

Keywords:Thermo-electric generatorPortable energy productionMeso-scale catalytic combustorPropane catalytic combustion

a b s t r a c t

The recent advances in miniaturized mechanical devices open exciting new opportunities for combus-tion, especially in the field of micro power generation, allowing the development of power-supply deviceswith high specific energy. The development of a device based on a catalytic combustor coupled with ther-moelectric modules is particularly appealing for combustion stability and safety. Furthermore, whenimplemented in micro/meso scale devices, catalytic combustion allows full utilization of hydrocarbonfuel’s high energy densities, but at notably lower operating temperatures than those typical of traditionalcombustion. These conditions are more suitable for coupling with conventional thermoelectric modulessince they prevent the modules’ degradation.In this work a novel catalytic meso-scale combustor fueled with propane/air mixture has been coupled

with two conventional thermoelectric modules. The wafer-like combustor is filled with commerciallyavailable catalytic pellets of alumina with Platinum (1% weight). Temperature measurements, in termsof point values and 2D distribution across the combustor surfaces, were carried out in different operatingconditions. Exhaust gases concentration measurements and pellet aging investigation were performed todetermine the combustor efficiency and stability. Starting from these results the combustor has beencoupled to bismuth telluride-based thermoelectric modules using a water cooled heat exchanger atthe cold side. The system obtained produces 9.86 W of electrical power reaching an overall efficiencyup to 2.85%: these results represent an improvement in portable-scale electrical power production fromhydrocarbon fuels state-of-art. Moreover, the voltage and current characteristics allow using such gener-ator for small portable devices power supplying.

! 2015 Elsevier Ltd. All rights reserved.

1. Introduction

During the last few years, the miniaturization of mechanicaland electromechanical engineering devices has received growingattention thanks to the increasing interest in the areas of micro-electronics, biomechanics, molecular biology, as well as thanks tothe progress made in microfabrication techniques [1]. Theadvances in miniaturized mechanical devices open exciting new

opportunities for combustion, especially in the field of micropower generation [2,3], allowing the development of power-supply devices with high specific energy (small size, low weight,long duration) [4,5]. Due to the small scale, effects of flame-wallinteraction and molecular diffusion become more significant inmicro and meso-scale combustion [2,4] when compared to tradi-tional combustion systems. These effects are still not completelyunderstood and require further investigation.

Even at 10% energy conversion efficiency, hydrocarbon fuels canprovide 10 times the energy density of the most advanced batteries[6–8]. Therefore, hydrocarbon-based devices as portable power

http://dx.doi.org/10.1016/j.apenergy.2015.10.0790306-2619/! 2015 Elsevier Ltd. All rights reserved.

⇑ Corresponding author.E-mail address: merotto@ieni.cnr.it (L. Merotto).

Applied Energy 162 (2016) 346–353

Contents lists available at ScienceDirect

Applied Energy

journal homepage: www.elsevier .com/ locate/apenergy

sources [7–9] can be considered as a suitable alternative to com-mon batteries. In this context, both homogeneous combustion[10–16] and catalytic reactors [17–22] are of major interest. Thesesystems are used either for hydrogen generation for fuel cells[23–27] or for direct conversion of thermal energy released viacombustion to electrical energy using thermoelectric or thermo-photovoltaic devices [28–30].

The major drawback of conventional homogeneous (gas-phase)combustion systems lies in the very high (>1500 "C) operatingtemperatures. These high temperatures greatly limit materialselection and require extensive combustor insulation. Furtherstrong limitations of these devices are strictly related to theirdimensions. As the size decreases, the surface area-to-volume ratioof the combustor increases with a subsequent increase of the heat-loss to heat-generation ratio. These strong losses in the smalldimension induce flame quenching [1–3], and are also responsiblefor an increase in pollutants emission such as CO and unburnedhydrocarbons. In addition, the smaller the volume, the shorterthe flow residence time. To overcome such limitations, catalyticcombustion seems to be the most suitable solution [11,31]. Whenimplemented in micro/meso scale devices, catalytic combustionallows full utilization of the high energy densities of hydrocarbonfuels, but at notably lower operating temperatures than those typ-ical of traditional combustion. Additionally, catalytic systems aregenerally easier to start, more robust to heat losses, and self-sustained at leaner fuel/air ratios [32–35].

When characterizing a catalytic combustor, catalyst degrada-tion needs to be addressed [36]. The main mechanisms for plat-inum catalyst degradation include particle agglomeration,platinum loss and redistribution, and poisonous effects due to con-taminants [37]. The catalyst degradation can be affected by theoperating conditions, including temperature, humidity, andsources of contamination of catalyst active sites. Some studies con-cerning platinum aging show that no great effect on Pt catalyst isdetected at automotive normal operating temperature (500 "C),with consequently no substantial change in overall system perfor-mance [38]. Despite the huge amount of literature on catalystdegradation, no results for tests performed in conditions similarto those addressed in this work are available in literature. There-fore, in the present work, degradation tests were carried out oncatalytic pellets.

In order to investigate the overall performance of a catalyticmeso-scale combustor, the analysis of exhaust gases is requiredto derive chemical efficiency. Fourier transform infrared spec-troscopy (FT-IR) has been widely applied for quantitative measure-ment of gas species concentrations of vehicles’s exhaust [39,40].Nevertheless, only a few works, such as [30], are reported in liter-ature regarding the application of the FT-IR technique to measuregases concentration at the exhaust of a meso-scale combustor.

Compared to fuel cells or other combustion-based systems,thermoelectric (TE) power generators are attractive for portablesystems due to their compactness, their reliability, their lowenergy consumption, and their high power densities. Consideringcommercial bismuth telluride-based TE modules, the operatingtemperature range is limited to 250 "C, which is incompatible withthe temperatures of a standard combustor. So, catalytic combus-tion is particularly well suited for TE power generation becauseof this relatively low temperature ceiling.

There are many examples of micro- and mesoscale thermoelec-tric power generators powered by catalytic combustion in avail-able literature [41–45]. A TE generator fabricated from siliconbonded to glass developed by Yoshida et al. [45] was able to pro-duce 184 mW of electrical power with an efficiency of 2.8% fromthe catalytic combustion of hydrogen. Norton and co-workers [5]reported for their integrated combustor-TE generator the produc-tion of 1 W maximum power and a thermal-to-electrical conver-

sion efficiency of 1.08% with hydrogen as the fuel. The group hasalso reported the generation of 0.45 W electrical power with pro-pane as the fuel at 0.42% conversion efficiency, using propanelower heating value LHV [27]. In recent years, Marton and co-workers [46] developed a butane-fueled thermoelectric generator(TEG) delivering 5.82 W maximum power with 2.53% conversionefficiency. Several efforts have been made by different authors toinvestigate the hot gases recirculation in order to reduce heatlosses and thus increase the efficiency of this kind of systems[47–49]. Nevertheless, the demonstration of a fuel-based TEG suit-able for portable power with energy density comparable to that ofa battery remains an open challenge.

In this work, a new premixed catalytic combustor has beendeveloped and characterized. The device has been designed forcoupling with thermoelectric modules for portable electricalpower production. The combustor is used as heat source in a sand-wich between two TE modules, in order to achieve a larger electri-cal power output, with I–V characteristics suitable for supplyingsmall electrical devices. Since the voltages obtained with smallthermoelectric devices are usually too low for direct utilization,this feature is one of the most important advantages offered bythe TEG presented in this work.

The meso-combustor developed in this work has notableadvantages when compared to similar solutions presented in theliterature [5–7]. The first advantage is the use of a low-cost, com-mercially available catalyst, with no need for ad hoc manufactur-ing. The second important advantage is that the use of such acatalyst allows obtaining the desired wall temperatures for ther-moelectric device coupling with no need for wall insulation fromhigher temperatures, thus resulting in smaller fuel consumptionand a more compact design. The commercial TE modules chosenare not specifically designed for power applications, but the refinedthermal control of the combustor allows producing up to 9.86 W ofelectrical power with a thermal to electrical conversion efficiencyof 2.85%, resulting in an improvement in portable-scale electricalpower converters based on thermoelectric technology and hydro-carbon fuels.

2. Materials and methods

A thermoelectric generator was designed coupling a catalyticpremixed meso-scale combustor and thermoelectric commercialmodules. The sketch of the TEG developed is reported in Fig. 1.Two TE modules are placed in a thermal chain consisting of the cat-alytic combustor surfaces (hot side) and two water-cooled heatsinks (cold side). In order to provide the most efficient contact atthe interfaces between the TE modules with the combustor surfaceand the heat sink, graphite sheets (100 lm thick) are used. Gra-phite was chosen for its capability for compensating the effectsof surface roughness and was preferred to thermal pastes for itshigh reproducibility and stability. The heat sinks are aluminumbulks homogeneously cooled by a 13 "C water flowing in the circuitat 6 l/min volume rate. The cooling system is arranged in order toprovide similar thermal conditions at both TEG sides. To ensure thethermal chain, the system is held together by two metal springsand four bolds and nuts. The compressive system allows obtaininga homogeneous pressure on the surfaces, minimizing thermallosses, compensating the increase of pressure on the TE modulesdue to the thermal expansion, and preserving the thermal chaincomponents’ correct alignment. In this work, 40 " 40 mm2 Fer-rotec commercial modules based on chalcogenides are used. Thesemodules are designed for cooling applications, so no data specificfor power generation are available. The operative limits below260 "C considered in this work are derived considering the tech-nologies used for thermoelectric elements soldering inside the

L. Merotto et al. / Applied Energy 162 (2016) 346–353 347

device. Moreover, each module has been tested in order to definethe conversion capability of the device, obtaining a response interms of Seebeck effect of 30 mV/K and an internal resistance, use-ful for maximum power at the matching load calculation, of3.94X. The power output is a function both of the load and ofthe DT applied to the module: using a DT = 150 "C, the maximumpower produced by a single module is in the order of 3.8 W corre-sponding to a V–I couple of 3.5 V and 1.1 A, respectively.

The catalytic premixed meso-scale combustor was designedand characterized to be coupled to the TE modules [50]. The inves-tigation was performed in order to assess the feasibility of themeso-scale combustor for the limits imposed by the modules used.Devices similar to the one presented in this work were proposed in[5], where the combustion channel is less than 1 mm high and thetemperature at the combustor walls is too high for a direct cou-pling with thermoelectrics, thus requiring an intermediate coolingstep. In the present work, the experimental conditions (e.g. com-bustion channel size, fuel/air mass flow rates, etc) were chosenin order to reach the desired temperature without any cooling sys-tem at the interface between combustor and TE modules.

In Fig. 2 a sketch of the combustor is shown. The geometry usedfor the combustion chamber was chosen in order to match the sizeof the TE devices. The aluminum-made combustor has a40 " 40 " 4 mm3 chamber filled with 155 commercial alumina

cylindrical pellets (r = 1.6 mm, h = 3.2 mm) covered with a thin cat-alytic film (platinum 1% weight). The pellets are placed in orderedlines so that the height of the channel for gas flowing is 0.8 mm.Propane and air are used as fuel and oxidizer, and the related massflow rates are measured and controlled by using thermal mass flowmeters (Bronkhorst El-Flow F-201CV). As C3H8/air mixture is notself-igniting in contrast with H2/air mixture, hydrogen assistedignition is performed.

The wall temperature distribution was measured with a FlirSystems Thermovision A40 thermo camera using the values fromK-type thermocouples placed on the combustor surface (A, B andC positions in Fig. 2) as absolute temperature reference for calibra-tion. The accuracy in point measurements is about 0.5% ±2 "C in thetemperature range considered (from #20 "C to +350 "C).

The overall chemical efficiency was estimated by means ofFourier transform infrared spectrometry (FT-IR) analysis of theexhaust gases. The exhaust gases were sent through a cut-off par-ticulate filter and a water trap to a Thermo Scientific Nicolet 6700FT-IR spectrometer equipped with a variable-pathlength heatedgas-cell (Gemini Mars series 6.4 M, internal volume of 0.75 l) posi-tioned inside the instrument. In order to prevent possible carbon-residuals condensation, both the transfer line and the gas-cell weremaintained at 393 K. To avoid exhaust gases dilution with ambientair, the suction mass flow was set to 6 Nl/min to be lower than theoutlet mass flow. The FT-IR spectra have a resolution of 0.5 cm#1.The background was obtained with N2 flow and subtracted fromthe spectra. Quantitative analyses of the gas concentrations wereperformed referring to a calibration based on mixtures of knowncompositions. The process chemical efficiency, gc, was calculatedusing the relationship:

gc ¼½CO2&

½CO2& þ ½CO& þ ½UHC&ð1Þ

where [UHC] is the concentration of the total unburnedhydrocarbons.

A peculiarity of thermoelectric technology is its long life withno need for maintenance. In order to consider the catalytic com-bustor a suitable heat source for the TEG, it was necessary to studythe combustor components aging effects under operative condi-tions. In particular, attention was focused on the pellets used. Aninvestigation of combustion processes effects on the degradationof the pellets catalytic coating was performed with metallographicanalyses by using a scanning electron microscope (SEM) (FEG Hita-chi SU-70) equipped with EDS and STEM detector.

The combustor was characterized at different operating condi-tions with the total mass flow rate ranging between 1.7 " 10#5

and 2.5 " 10#5 kg/s varying the equivalence ratio, U, in the range0.8–1.2 by changing both the propane and the air mass flow rate.

The TEG performance investigation was carried out at differentoperating conditions of the catalytic combustor with the total massflow rate ranging between 6.3 " 10#5 and 14.6 " 10#5 kg/s, keep-ing U at stoichiometric condition. Control of the final DT applied

Fig. 1. Sketch of the thermoelectrical generator obtained by coupling the combustor with two commercial thermoelectic modules.

Fig. 2. Sketch of the meso-scale combustor used in this work: the positions of thethermocouples used to measure the combustor wall temperature are indicate withA, B, C and D.

348 L. Merotto et al. / Applied Energy 162 (2016) 346–353

to the TE modules was achieved varying the inlet total mass flowrate, increasing in this way the hot side temperature. The cold sidetemperature remained constant during the characterization pro-cess, due to the power removed by the water cooled circuit, prov-ing that the good thermal coupling of the TEG’s different elementsallowed obtaining a good control of the stability at the cold side ofthe system. The TEG electrical output was studied in terms of volt-age and power produced for each DT applied. The I–V and I–Wcharacteristics were measured at thermally steady conditions byconnecting the TEG output to external loads having increasing val-ues from short-circuit up to open-circuit.

The data obtained from combustor and TEG characterizationallowed calculating the electrical conversion efficiency for eachmodule and for the overall device.

3. Results and discussion

The first step of the combustor characterization was assessingthe total mass flow rate in order to verify the temperature on theexternal walls at different conditions. Preliminary tests showedthat the lower combustion stability limit is at U = 0.8, for2.5 " 10#5 kg/s mass flow rate.

In Fig. 3, the temperature values measured in the four positionsversus U for three different total mass flow rates (1.7 " 10#5 kg/s,2.1 " 10#5 kg/s and 2.5 " 10#5 kg/s) are shown. The uncertainty inthe measured temperatures is approximately 6.67% in the worstcase. Using 50.35 " 106 J/kg for the propane calorific value, thermalpower values between 62 and 95W were obtained for total massflow rates ranging from 1.7 " 10#5 kg/s to 2.5 " 10#5 kg/s,respectively.

As it can be noted, wall temperature increases with U up toU = 1, then remains approximately constant for higher values.The behavior is similar for the three flow rates investigated at allpositions. The temperature slightly decreases moving downwardfrom the inlet at fixed flow rate, and increases monotonically withthe flow rate. Exhaust gases temperature values are typicallyslightly above 300 "C, which is higher than the walls temperaturesin each condition.

The 2D temperature mapping on the combustor externalwalls is reported for different flow rates in Fig. 4. Only the actualcombustion chamber surface has been considered. The arrowsrefer to the mixture inlet and outlet. The actual inlet in the com-bustor is placed horizontally (45" rotation with respect to theFigure).

Fig. 3. Temperature vs equivalence ratio at four different locations on the combustor wall and at three different mass flow rates.

Fig. 4. Temperature distribution on the combustor chamber external walls at three different mass flow rates. Arrows show gas inlet and outlet direction. Circles highlight theaverage temperature obtained in each case. Temperatures are given in Celsius degrees.

L. Merotto et al. / Applied Energy 162 (2016) 346–353 349

The thermocamera images were calibrated using the thermo-couple data previously described. Investigation performed on bothcombustor sides showed no significant differences. The resultsprove that the temperature on the external surface of the wall issatisfactorily uniform, thus confirming that the meso combustoris a suitable hot source for the TEG. It can be noticed that theobtained temperature is lower than the typical propane catalyticreaction temperature (700–800 "C). This can be explained takinginto account the heat dissipation occurring at the combustor sur-face due to the thermal exchange with ambient air. Consideringthe higher temperatures obtained at the combustor outlet

(Fig. 3), it is reasonable to assume that the internal temperaturematches the catalytic reaction temperature found in literature.

The FT-IR spectra of the gases collected at the exhaust of thepremixed meso-scale combustor were analyzed to determine theircomposition. After calibration, it was possible to perform quantita-tive analyses with an estimated accuracy below 1% for FT-IRmeasurements.

In Fig. 5 propane concentration versus U is reported for thethree mass flow rates investigated. In all cases, propane concentra-tion profile shows a minimum at U = 1 as expected. For lower andhigher equivalence ratios, higher unburned propane was found atthe combustor outlet. At U = 1, the unburned propane increaseswith the decrease of the mass flow rate, ranging from about490 ppm for 2.5 " 10#5 kg/s mass flow rate to 940 ppm for1.7 " 10#5 kg/s. The measurements error is small, with the uncer-tainty in C3H8 concentration being about 7.5% in the worst case.

CO2 and CO concentrations were also measured. In both casesthey increase with U, reaching a plateau above U = 1. The samebehavior was obtained changing the mass flow rate values.

Taking into account the concentration measurements, it is pos-sible to evaluate the chemical efficiency given by Eq. (1) wherepropane is the only unburned hydrocarbon detected. In Fig. 6, effi-ciency vs. equivalence ratio is shown for all the tested flow rates.The uncertainty in the efficiency value is 0.5% in the worst case.A slight dependence on U was observed, showing a maximum atthe stoichiometric condition where about 96% efficiency isreached. For higher and lower U values, efficiencies around92–94% are obtained. Similar values of this chemical efficiencyhave been reported in literature [5], referring to a different com-bustor fueled with the same mixture at higher flow rates.

The results reported for the catalytic combustor characteriza-tion prove that the system thermal control and homogeneity aresuitable for coupling the combustor with TE modules obtaininggood chemical efficiency.

As complementary analyses to evaluate the combustor life span,tests on pellet degradation were performed. The pellets underwent200 h operating combustion conditions, and the evolution of theactive phase was verified with SEM analyses. It is reasonable toexpect that the degradation takes place at typical temperature val-ues of catalytic combustion (e-g. 700–800 "C) [34]. In these condi-tions, coking, agglomeration and contamination effects of thecatalytic pellets are likely to occur, which could affect the perfor-mance of the overall system.

SEM images (backscattered electrons) for fresh and used (200 h)pellets surfaces are shown in Fig. 7. It can be noted that used pel-lets showwhite agglomerates spread over the surface which do notappear on the fresh pellets surface. EDX analyses suggest that theseagglomerates consist of Platinum, while the main material isalmost entirely alumina.

Fig. 5. Combustor exhaust gases analysis: unburned propane concentration vs.equivalence ratio for three different mass flow rates.

Fig. 6. Chemical efficiency versus equivalence ratio for the mass flow ratesinvestigated.

Fig. 7. SEM backscattered micrographs of fresh (left) and 200 h aged (right) pellet base surface.

350 L. Merotto et al. / Applied Energy 162 (2016) 346–353

Pt distribution mapping of the pellets’ surface (Fig. 8) highlightsthat very small Pt spots are spread on the fresh pellets’ surface,while the used pellets’ surface shows fewer but much larger Ptstructures reaching a size of 2–3 lm2. Thus, the pellets morphol-ogy evolves in operating time due to the combustion temperatureand the catalyzer aggregates forming particles which could affectits efficiency in supporting the combustion. However, after 200 hthe combustor performances seemed not to be affected by this evo-lution and the combustor maintains its characteristics.

In Figs. 9 and 10, the results of the characterization of the cou-pled system, catalytic combustor and TE modules, in terms of elec-trical output of the TEG are reported. The output voltage (V) andthe electrical power (W), respectively, are plotted as a function ofthe current produced for different DT applied. The maximumpower delivered by the device ranges between 5 and 10 W forDT values ranging between 100 and 200 "C. Correspondingly, thevoltage drop across the load is in the range 7.5–10 V. The valuesobtained are consistent with the preliminary characterization atdifferent thermal conditions performed on the commercial mod-ules used.

The experiments on the TEG were performed for different total(propane + air mixture) mass flow rates resulting in different DTbetween the hot and cold sides. Experimental parameters for thedifferent tests and the results obtained are summarized in Table 1.Each total mass flow rate condition was obtained keeping U con-stant at stoichiometric conditions and by varying the propaneand air mass flow rates.

It can be noted that all the mass flow rate conditions used forTEG characterization are higher than those previously reportedfor the combustor alone. This is because when the combustor iscoupled with the TE modules and the cooling system, the desiredtemperature at the combustor surface (up to 250 "C at the hot side)is obtained for higher flow rates due to the increased thermalexchange. The minimum mass flow rate (6.3 " 10#5 kg/s) corre-sponds to the limit under which the combustion is no longer sus-tained, while the maximum mass flow rate (14.6 " 10#5 kg/s) isthe limit over which no further DT increase is obtained with thecooling system used.

The power provided to and the power delivered from the deviceare plotted in Fig. 11 as a function of total mass flow rate. It can benoted that the obtained power curve slope tends to decrease forhigher mass flow rates, indicating that the system is approachingits maximum performance at the given cooling conditions. Nohigher mass flow rate was tested in this work because at14.6 " 10#5 a combustor surface temperature of 255 "C wasreached, which is the maximum allowed for the TE modules used.The uncertainty on the experimental point is ±2 W for the powerprovided to the system and ±0.8 for the power obtained.

In order to obtain the device thermal to electrical conversionefficiency, the power provided to the system was calculated as

Fig. 8. EDX mapping of Pt distribution on fresh (left) and 200 h aged (right) pellets surface.

Fig. 9. Power delivered by the TEG vs. current.

Fig. 10. Voltage of TEG vs. current.

Table 1Maximum TE power obtained, power provided, and conversion efficiency at differentmass flow rate and DT values.

Mixture massflow rate (kg/s)

DT("C)

Thot("C)

Powerprovided(W)

Powerobtained(W)

Conversionefficiency (%)

6.3 " 10#5 100 179 178.71 5.06 2.838.4 " 10#5 130 207 238.29 6.78 2.8510.3 " 10#5 155 233 297.86 8.24 2.7712.6 " 10#5 190 245 357.43 9.24 2.5914.6 " 10#5 200 258 417.00 9.86 2.36

L. Merotto et al. / Applied Energy 162 (2016) 346–353 351

the propane mass delivered in each testing case times its lowerheating value (LHV). Fig. 12 shows the thermal to electrical conver-sion efficiency obtained for the different mass flow rate valuestested. The maximum efficiency obtained is around 2.85%. Similarvalues are obtained for mass flow rates values up to10.3 " 10#5 kg/s, while for higher mass flow rates the efficiencytends to decrease. This behavior can be explained considering theeffect of increased mass flow rate on exhaust concentrationdescribed before: unburnt propane concentration increases withincreasing mass flow rates due to the reduced residence time. This

leads to a lower propane conversion efficiency which negativelyaffects the thermal to electrical conversion.

A comparison with similar devices described in literature isshown in Table 2, in which the overall power produced and thethermal to electrical conversion efficiency in different TEGs basedon catalytic combustion are shown. The TEG presented in this workrepresents an improvement in the conversion efficiency comparedto other published works.

Future developments of the TEG device presented in this workwill include an upgrade of the ignition system with inlet gases pre-heated through exhaust gases recycling, and preheated inlet gasesreplacing the hydrogen-assisted ignition.

4. Concluding remarks

A novel premixed catalytic meso-scale combustor fueled withpropane/air was developed to be coupled with a thermo-electricdevice for portable electric applications. Due to the need to usecommercially-available thermoelectric devices, the limitation oftemperature at 250 "C was a stringent requirement to match.Therefore, the proper mass flow rate value was found inside thestability limits which produced reasonable values of temperatureon the combustor surfaces. In order to investigate the temperaturedistribution on the combustor surfaces, temperature measure-ments were carried out by using an IR camera. A uniform distribu-tion on both sides was obtained, assuring thermo-electricconversion optimization.

Quantitative analysis of the gas-species concentration of theexhaust was performed by FT-IR measurements. From these mea-surements, the propane conversion efficiency values were derivedin the different experimental conditions under study. It was foundthat reasonable high efficiency of about 96% was obtained at thestoichiometric condition.

Considering the catalytic pellets degradation during combus-tion, a change of the distribution of Pt on the alumina surface wasobserved, resulting in bigger and more complex structures. No sig-nificant effects of the coating degradation on the overall combustorperformance were observed after 200 h. These results suggest thatthe combustor proposed is suitable for thermo-electric conversionapplication. The controlled combustion conditions and the low con-sumption of the Pt layer suggest a longer expected lifetime for a cat-alytic combustor in comparison to a traditional combustor. Thisaspect couples the thermoelectric device advantage in terms of longlife without maintenance: the generator has no moving parts anddemonstrates long-lasting operative stability.

The catalytic combustor was integrated with TE modules to pro-duce 9.86 W of electrical power with a thermal to electrical con-version efficiency of 2.85%. The proposed device represents anotable improvement in portable-scale electrical power produc-tion from hydrocarbon fuels, with an increase in the conversionefficiency compared to other published works.

Future perspectives include reducing the system scale, andexploiting the hot gases recirculation in order to reduce the heatlosses and further increase the thermal to electrical conversionefficiency. The downscale of the dimension is going to focus onmaintaining the highest possible level of power production butwith a specific care for the voltage level obtained, in order toensure both ranges are suitable for power supplying of portableelectronic devices.

Acknowledgements

The authors would like to acknowledge the financial support forthe project INTEGRATE in the framework of CNR-Regione Lombar-dia program, the technical assistance of Mr Fantin, and Ms. PaulaAlbin Reynolds for editing contribution.

Fig. 11. Power provided to the TEG (based on LHV) and power obtained vs. mixturemass flow rate.

Fig. 12. Efficiency of thermal to electric power conversion vs. mixture mass flowrate.

Table 2Comparison of the TE power generation obtained in this work with those of otherpublished works.

Source Fuel Overall powerproduced (W)

Conversionefficiency (%)

Federici [27] Propane 0.45 0.42Norton [5] Hydrogen 1 1.08Marton [46] Butane 5.82 2.53Yoshida [45] Hydrogen 0.18 2.8This work Propane 9.86 2.85

352 L. Merotto et al. / Applied Energy 162 (2016) 346–353

References

[1] Vican JGBF, Gajdeczko BF, Dryer FL, Milius DL, Aksay IA, Yetter RA.Development of a microreactor as a thermal source formicroelectromechanical systems power generation. Proc Combust Inst2002;29(1):909–16.

[2] Maruta K. Micro and mesoscale combustion. Proc Combust Inst 2011;33(1):125–50.

[3] Yang WM, Chou SK, Shu C, Xue H, Li ZW, Li DT, et al. Microscale combustionresearch for application to micro thermophotovoltaic systems. Energy ConversManage 2003;44(16):2625–34.

[4] Fernandez-Pello AC. Micropower generation using combustion: issues andapproaches. Proc Combust Inst 2002;29:883–99.

[5] Norton DG, Voit KW, Brüggemann T, Vlachos DG, Wetzel ED. Portable powergeneration via integrated catalytic microcombustion-thermoelectric devices.In: 24th Army Science Conference, Orlando, FL, 2004.

[6] Hua J, Wu M, Kumar K. Numerical simulation of the combustion of hydrogen–air mixture in micro-scaled chambers Part II: CFD analysis for a micro-combustor. Chem Eng Sci 2005;60(13):3507–15;Patil AS, Dubois TG, Sifer N, Bostic E, Gardner K, Quah M, et al. Portable fuel cellsystems for America’s army: technology transition to the field. J Power Sources2004;136(2):220–5.

[7] Yadav S, Sharma P, Yamasani P, Minaev S, Kumar S. A prototype micro-thermoelectric power generator for micro-electromechanical systems. ApplPhys Lett 2014;104(12):123903.

[8] Dunn-Rankin D, Leal EM, Walther DC. Personal power systems. Prog EnergyCombust Sci 2005;31:422–65.

[9] Choi W, Kwon S, Dong Shin H. Combustion characteristics of hydrogen–airpremixed gas in a sub-millimeter scale catalytic combustor. Int J HydrogenEnergy 2008;33(9):2400–8.

[10] Norton DG, Wetzel ED, Vlachos DG. Fabrication of single-channel catalyticmicroburners: effect of confinement on the oxidation of hydrogen/airmixtures. Ind Eng Chem Res 2004;43:4833–40.

[11] Wierzbicki TA, Lee IC, Gupta AK. Combustion of propane with Pt and Rhcatalysts in a meso-scale heat recirculating combustor. Appl Energy2014;130:350–6.

[12] Pattekar AV, Kothare MV. A radial microfluidic fuel processor. J Power Sources2005;147:116–27.

[13] Norton DG, Vlachos DG. A CFD study of propane/air microflame stability.Combust Flame 2004;138:97–107.

[14] Kaisare NS, Vlachos DG. Optimal reactor dimensions for homogeneouscombustion in small channels. Catal Today 2007;120:96–106.

[15] Miesse CM, Masel RI, Jensen CD, Shannon MA, Short M. Submillimeter-scalecombustion. AIChE J 2004;50(12):3206–14.

[16] Kaisare NS, Deshmukh SR, Vlachos DG. Stability and performance of catalysticmicroreactors: simulations of propane catalytic combustion on Pt. Chem EngSci 2008;63:1098–116.

[17] Norton DG, Wetzel ED, Vlachos DG. Thermal management in catalyticmicroreactors. Ind Eng Chem Res 2006;45:76–84.

[18] Ouyang X, Bednarova L, Besser RS, Ho P. Preferential oxidation (PrOx) in a thin-film catalytic microreactor: advantages and limitations. AIChE J 2005;51(6):1758–72.

[19] Rebrov EV, De Croon MHJM, Schouten JC. A kinetic study of ammoniaoxidation on a Pt catalyst in the explosive region in a microstructured reactor/heat-exchanger. Chem Eng Res Des 2003;81(7):744–52.

[20] Deshmukh SR, Vlachos DG. A reduced mechanism for one-step rateexpressions for catalytic combustion of small alkanes on noble metals.Combust Flame 2007;149:366–83.

[21] Saracco G, Veldsink JW, Versteeg GF, van Swaaij WP. Catalytic combustion ofpropane in a membrane reactor with separate feed of reactants—II. Operationin presence of trans-membrane pressure gradients. Chem Eng Sci 1995;50(17):2833–41.

[22] Deshmukh SR, Mhadeshwar AB, Lebedeva MI, Vlachos DG. From densityfunctional theory to microchemical device homogenization: model predictionof hydrogen production for portable fuel cells. Int J Multiscale Comput Eng2004;2(2).

[23] Ganley JC, Seebauer EG, Masel RI. Development of a microreactor for theproduction of hydrogen from ammonia. J Power Sources 2004;137(1):53–61.

[24] Deshmukh SR, Vlachos DG. Effect of flow configuration on the operation ofcoupled combustor/reformer microdevices for hydrogen production. ChemEng Sci 2005;60(21):5718–28.

[25] Kolios G, Glöckler B, Gritsch A, Morillo A, Eigenberger G. Heat-integratedreactor concepts for hydrogen production by methane steam reforming. FuelCells 2005;5(1):52–65.

[26] Norton D, Deshmukh SR, Wetzel ED, Vlachos DG. Downsizing chemicalprocesses for portable hydrogen production. ACS symposium series, vol.914. Oxford University Press; 2005. p. 179–93.

[27] Federici JA, Norton DG, Bruggeman T, Voit KW. Catalytic Microcombustor withintegrated thermoelectric elemenst for portable production. J Power Sources2006;161:1469–78.

[28] Ahn J, Eastwood C, Sitzki L, Ronney PD. Gas-phase and catalytic combustion inheat-recirculating burners. Proc Combust Inst 2005;30(2):2463–72.

[29] Zhang S, Yuan XZ, Hin JNC, Wang H, Friedrich KA, Schulze M. A review ofplatinum-based catalyst layer degradation in proton exchange membrane fuelcells. J Power Sources 2009;194(2):588–600.

[30] Wu M, Yetter RA, Yang V. Combustion in meso-scale vortex combustors:experimental characterization. In: 42nd AIAA Aerospace Sciences Meeting andExhibit, 5–8 Jan. 2004, Reno, Nevada. AIAA 2004-980, 2004.

[31] Neyestanaki AK, Kumar N, Lindfors L-E. Catalytic combustion of propane andnatural gas over Cu and Pd modified ZSM zeolite catalysts. Appl Catal B 1995;7(1):95–111.

[32] Vijayan V, Gupta AK. Combustion and heat transfer at meso-scale with thermalrecuperation. J Appl Energy 2010;87(8):2628–39.

[33] Vijayan V, Gupta AK. Flame dynamics of a meso-scale heat recirculatingcombustor. J Appl Energy 2010;87(12):3718–28.

[34] Vijayan V, Gupta AK. Thermal performance of a meso scale liquid fuelcombustor. J Appl Energy 2011;88(7):2335–43.

[35] Shirsat V, Gupta AK. Performance characteristics of methanol and kerosenefuelled meso-scale heat recirculating combustors. J Appl Energy2011;88:5069–82.

[36] Yazawa Y, Takagi N, Yoshida H, Komai SI, Satsuma A, Tanaka T, et al. Thesupport effect on propane combustion over platinum catalyst: control of theoxidation-resistance of platinum by the acid strength of support materials.Appl Catal A 2002;233(1):103–12.

[37] González-Velasco JR, Botas JA, Ferret R, Pilar González-Marcos M, Marc JL,Gutiérrez-Ortiz MA. Thermal aging of Pd/Pt/Rh automotive catalysts under acycled oxidizing–reducing environment. Catal Today 2000;59(3):395–402.

[38] Mayrhofer KJ, Meier JC, Ashton SJ, Wiberg GK, Kraus F, Hanzlik M, et al. Fuelcell catalyst degradation on the nanoscale. Electrochem Commun 2008;10(8):1144–7.

[39] Hanst PL. Infrared spectroscopy and infrared lasers in air pollution researchand monitoring. Appl Spectrosc 1970;24(2):161–74.

[40] Baum MM, Kiyomiya ES, Kumar S, Lappas AM, Kapinus VA, Lord HC.Multicomponent remote sensing of vehicle exhaust by dispersive absorptionspectroscopy. 2. Direct on-road ammonia measurements. Environ Sci Technol2001;35(18):3735–41.

[41] Aranguren P, Astrain D, Rodríguez A, Martínez A. Experimental investigation ofthe applicability of a thermoelectric generator to recover waste heat from acombustion chamber. Appl Energy 2015;152:121–30.

[42] Chou SK, Yang WM, Chua KJ, Li J, Zhang KL. Development of micro powergenerators – a review. Appl Energy 2011;88(1):1–16.

[43] Kinsella CE, O’Shaughnessy SM, Deasy MJ, Duffy M, Robinson AJ. Batterycharging considerations in small scale electricity generation from athermoelectric module. Appl Energy 2014;114:80–90.

[44] Xiao H, Qiu K, Gou X, Ou Q. A flameless catalytic combustion-basedthermoelectric generator for powering electronic instruments on gaspipelines. Appl Energy 2013;112:1161–5.

[45] Yoshida K, Tanaka S, Tomonari S, Satoh D, Esashi M. High-energy densityminiature thermoelectric generator using catalytic combustion.Microelectromech Syst, J 2006;15(1):195–203.

[46] Marton CH, Haldeman GS, Jensen KF. Portable thermoelectric power generatorbased on a microfabricated silicon combustor with low resistance to flow. IndEng Chem Res 2011;50(14):8468–75.

[47] Shirsat V, Gupta AK. A review of progress in heat recirculating meso-scalecombustors. Appl Energy 2011;88(12):4294–309.

[48] Tolmachoff Erik D, Allmon William, Mike Waits C. Analysis of a highthroughput n-dodecane fueled heterogeneous/homogeneous parallel platemicroreactor for portable power conversion. Appl Energy 2014;128:111–8.

[49] Vijayan V, Gupta AK. Combustion and heat transfer at meso-scale with thermalrecuperation. Appl Energy 2010;87(8):2628–39.

[50] Merotto L, Dondè R, De Iuliis S. Study of the performance of a catalyticpremixed meso-scale burner. In: Ninth Mediterranean CombustionSymposium, Rhodes, Greece, 7–11 June 2015.

L. Merotto et al. / Applied Energy 162 (2016) 346–353 353