Experimental Analysis of the Influence of Inert Nano-additives upon Combustion of Diesel Sprays

Recent experimental studies report that addition of nanoparticles (tipically metal oxides nanopowders, such as Al2O3) can somehow modify the ignition mechanism of liquid fuels, probably influencing and accelerating the thermal exchange process between the fuel droplets and the surrounding air. In this paper it has been experimentally analised a stationary Diesel spray, produced by a gas assisted atomizer, preliminarily comparing the behaviour of the flame generated using the unseeded fuel with that one of the same fuel addit ivated with Al2O3 nanoparticles. Results obtained through light emission analysis are encouraging, in the sense that the presence of a small nano-additive quantity (0.1% in volume) seems to produce effects similar to a shift towards higher flame ventilation, with a possible positive consequence upon combustion stability and efficiency.


Introduction
Nanostructured materials represent nowadays a wide and probably still largely unexp lored field of potential applications [1]. In fact, this is a research topic in high and rapid develop ment, both at a basic level and under the point of view of possible practical applications, leaving large space for a thorough scientific analysis, which requires with no doubt long time for u ltimate conclusions.
The potential application fields of nanoparticles are very wide and, probably, not yet fully understood: fro m energetics (sensors, propulsion, reduction of environmental impact) to chemistry (catalysis, additives) to med icine (diagnostics) and o f co u rs e to material s cien ce. A lso th e p ro b lem o f nanoparticles impact on human health is a rather unexp lored field. A mong the possible applications, recent experimental studies [2] repo rt that add ition o f nanopart icles (typically metal o xides nanopowders, such as Al2O3) can so mehow mod ify the ign it ion mechan is m of liqu id fuels, probab ly influencing and accelerating the thermal exchange process between the fuel droplets and the surrounding air. In fact, the added n anopart icles can be s mall enough to app roach mo lecular d imensions, modify ing the physical properties of th e liq u id fu el (s uch as thermal co n d uctiv ity , mass diffusivity and radiative heat transfer) and at the same time in creas ing the su rface-area-to -vo lu me ratio of th e fu el droplets allowing more fuel to be in contact with the oxidant.
These studies are focussed on the analysis of the ignition probability of additivated fuel single droplets imp inging on a hot plate and it has been observed that ignition temperature of fuel mixtures containing nanoparticles is lower than that of pure Diesel fuel. In fact, ignition delay and ignition temperature are critical parameters to characterize the performance of a Diesel fuel, influencing both emission levels and efficiency of the co mbustion process. Moreover, it has been observed [3,4] that the use of nanoparticles in rocket propulsion can enhance the engine performance under the point of view of specific thrust.
Anyway, no data have been yet reported about the possible influence of nanoparticles addition upon the combustion features of a Diesel spray. Therefore, in this paper it has been experimentally analysed a stationary Diesel spray, produced by a gas assisted atomiser, preliminarily co mparing the behaviour of the flame generated using the unseeded fuel with that one of the same fuel seeded with A l2O3 nanoparticles.

Experimental Set-up
As previously outlined, the core of the apparatus consists of a gas assisted atomiser consisting of a capillary tube (0.3 mm inner diameter, 0.8 mm outer d iameter) that lies in an opening of 1 mm d iameter at the spray top, thus creating an annular gap. The liquid fuel flows under the action of a syringe pump (New Era Pu mp Systems Inc., Model NE-1000) inside the capillary at a flo w rate of 1 ml/ min. The dispersion gas (air in this applicat ion) flows inside the annular gap. The capillary ends a little downstream the opening, giving rise to an atmospheric spray flame ignited by a premixed methane-air flame which is concentrically arranged around the nozzle exit . This sustaining flame was kept just slightly rich, thus originating an overventilated cone in the presence of the o xidant. The gas flow rates are measured and controlled by thermal mass flow meters (Bronkhorst High Tech. The Netherlands). The apparatus is designed in order to achieve, through the longitudinal displacement of the dispersion gas injector, a variation of the outflow area, fro m zero to the maximu m value. The co mbustion process has been confined inside a cylindrical quart z chamber (200 mm diameter), allo wing fu ll optical access, and the burned gases are conveyed in a hood connected with a sucking pump. The hood is equipped with a fibreglass filter (Whatman GF/A) placed rather far away fro m the flame tip, thus trapping the nanoparticles without been affected by hot gases.
A schematic v iew of the whole apparatus is reported in Fig.  1.
An easy way to comply with the paper Tab. 1 reports the main characteristics of the Diesel fuel used for the experimental analysis: as it can be seen, it is a fuel very close to the commercial ones used for automotive applications. A rather similar apparatus was used to characterise liquid fuels spray flames [5] and for controlled nano-particles synthesis by flame spray pyrolysis [6][7][8][9][10]. The Al 2 O 3 nanoparticles used are produced by Nanostructured and Amorphous Materials Inc. (nominal average diameter=10 n m) with 99% cert ified purity. To obtain the alumina added fuel, the nanoparticles have been weighed and subsequently mixed in the fuel with different sonication cycles, ensuring the highest dispersion degree. Al 2 O 3 has been selected because it can be considered as inert with respect to the combustion reaction, putting into evidence only a possible physical interaction with the various phenomena involved in spray combustion. Moreover, Al 2 O 3 nanoparticles are rather easily available at a relative low cost, taking into account also the very low quantity of nanoparticles dispersed in the liquid fuel. In fact, for the preliminary analysis described in th is paper, a volu metric percentage of 0.1% of nanopowders dispersed in the liquid fuel has been used. The percentage of the nanoparticles dispersed in the fuel has been selected on the basis of the results reported in [2] as the mimimu m value to obtain noticeable influence upon the ignit ion probability of the Diesel fuel. Therefore, this value has been kept fixed during the experimental measurements.
The fluid dynamic behaviour of the atomiser in isothermal conditions has been investigated through Phase Doppler Anemo metry (PDA).
The preliminary co mparison of the flame spray behaviour has been performed by emission spectroscopy and burned gases sampling and co mposition analysis. As for spectroscopy, it has been used a Tracor Northern apparatus equipped with a low resolution spectrograph. The radiation was collected by a quartz fibre through a yellow band-pass filter (597±30 n m): in fact, th is spectral region is the mo re interesting under the point of view of unburned species formation. These lead to soot formation and then to black body emission. The flame region optically exp lored was roughly corresponding to the upper third of the flame plu me. Burned gases have been sampled in the exhaust hood and the composition (NO x , CO, SO x , O 2 , CO 2 , excess air) has been characterised by an electrochemical analyser (Eurotron Greenline mod. M K2).

Results and Discussion
The experimental analysis has been performed upon the stationary spray, operated at atmospheric pressure.
In order to investigate the dimensional distribution of the generated droplets as a function of dispersion gas flow rate and gas to liquid mass ratio, the behaviour of the spray jet has been initially characterised in isothermal (co ld) conditions, both through visualizat ions and PDA (Phase Doppler Anemo metry). Figure 2 summarizes the most representative obtained results. Particularly, Fig. 2A shows the comparison of the counter mean d iameter (CMD) for the water spray and the n-hexane spray at 5 mm fro m the burner exit by varying the oxidant flow. For flows of o xidant higher than 2 l/ min the size of the droplet is almost constant, particularly for n-hexane, wh ich behaviour can be considered representative of hydrocarbon fuel. N-hexane fuel produces smaller droplets than water (about 11 μm against 19 μm). The radial profile of CMD was also investigated. As an examp le Fig.  2B shows the CMD rad ial profile for the water spray at 50 mm fro m the exit nozzle. The diameter o f the droplet is almost constant with a value of about 13 μm.
It can be concluded that the "cold" spray is formed by droplets of about 15-20 m average size, with operating conditions almost constant inside a wide range of ato mising air flow rate.
After the preliminary characterization of the spray behaviour, a qualitative imaging of the generated flame has been performed varying the operating conditions (i.e. main ly the air flow rate). Fig. 3 reports images of the spray flame obtained burning the pure fuel at a fixed flo w rate (input thermal power of the spray flame=700 W) varying the atomising air flow rate. It can be observed that a decrease of the air flow rate induces a progressive format ion of intense yellow coloured co mbustion regions due to higher particulate format ion fro m unburned species.  Figure 2. A comparison of the counter mean diameter (CMD) for water and n-hexane spray at 5 mm from the exit nozzle -B) radial profile of the CMD for the water spray at 50 mm from the exit nozzle Temperature measurements along a radial profile using a B-type thermocouple (200 µm hot junction dimension) have been carried out to analyse the thermal behaviour of the premixed sustaining flame. Fig. 4 reports the results obtained corrected for the radiat ive losses, as reported in [11].
The profile at lo wer height (blue, z=0mm) corresponds to a minimu m distance (about 1 mm) fro m the burner surface, which was the lowest accessible by the thermocouple arm. The profile has a hollow shape, owing to the receptacle of the spray nozzle, placed in the middle, then in the corrispondence of the bottom part of the spray cone (r/ R ≈ 0.8-1). Moving upwards the temperature p rofiles tend to close, reaching at the centre about 900°C just at 3 mm height (pale b lue). Here therefore, is where the nebulized spray is ignited (co mpare Fig. 3 ).    5 reports the comparison of the flame light emission (obtained as previously described) in the case of pure fuel and alumina loaded fuel, as a function of the ato mising air flow rate. For this measurement, and in general for all measurements of this kind, an integration t ime of 18 s was set, thus intrinsecally averaging short term fluctuations of the lu minosity. It can be clearly observed an effect due to nanoparticles presence, giving rise to a general decrease of emission intensity from the flame, with a higher influence (up to 50% of intensity decrease) at lower values of air flow rate, that is just in the more crit ical flame conditions under the ventilation and nebulisation point of view. At higher values of air flo w rate, the influence almost disappears and the curves tend to coincide. This, of course, does not mean that the effect of nanoparticles ceases, but simply that the effect is no longer dectectable with this optical technique.
It is important to stress that the result of Fig. 5 is purely qualitative and strongly dependent on the particular flame structure under investigation. For instance, just a little change in the nozzle arrangement could largely alter the spray features and consequently the flame shape and lu minosity in its absolute values, especially at the tip. The only mean ingful point here is the relat ive position of the two curves: the one laying below is expected to be originating fro m the flame exh ibit ing better combustion, as already observed in [12,13]. However, to support the results of Fig. 5, two couples of points (1.95 and 2.2 Nl/ min air) were tested with sets of 20 measurements each. All measurements were then averaged over the full set resulting in four experimental points covering a rather long measuring time. The lu minosity of the seeded flame resulted steadily below the one of the pure fuel.
In order to confirm the possible positive influence of nanoparticles presence upon the combustion process, burned gases have been sampled and analysed. In this case, CO emission emerged as the more significant result (NO x formation being almost negligib le and indistinguishable in the two cases). Fig. 6 su mmarizes the measurements obtained as a function of the atomising air flow rate: the presence of nanopowders in the fuel contributes to lower up to 7% the CO emission with respect to the pure fuel flame, confirming at least under a macroscopic point of view the improvement of combustion features due to nanoparticles addition.

Conclusions
The preliminary work performed suggests that the addition of nanoparticles to Diesel fuel (also at low concentration, i.e. 0.1% vol) can improve the co mbustion features of the spray flame, g iving rise to lower CO emission levels and simu lating a sort of higher-ventilation of the flame with respect to the unseeded spray, at identical air feeding conditions. Future experimental work is foreseen in order to deepen the results emerged till now, especially for sensitivity analysis of the spray flame to nanoparticles concentration in the fuel, typology and characteristic dimension of nanoadditives used, as a function of the operating conditions of the experimental set-up. Moreover, a theoretical study has to be performed in order to understand the influence of nanoparticles addition upon the combustion process, reaching a model of the involved phenomena.