In what follows the relevant literature on the present study is briefly reviewed.
Most of existing studies relevant to the impinging jet deal with heat and mass transfer of a single jet impinging onto a flat plate, emphasizing the high heat transfer efficiency of the jet. According to the flow characteristics of free jets, Mcnaughton and Sinclair [3] identified four main types of jet in their experimental study and classified the jet by its Reynolds number Rej: (1) dissipated-laminar jet for Rej<300, (2) fully laminar jet for 300<Rej<1000, (3) semi-turbulent jet for 1000< Rej<3000, and (4) fully turbulent jet for Rej>3000. As illustrated in Fig. 1.2, the flow in a jet impinging vertically onto a plate can be divided into three regions: (1) free jet region:
near the nozzle the jet flow mainly moves in the axial direction and is not affected to a noticeable degree by the presence of the impingement surface, (2) stagnation region:
which is located between the free jet region and the wall jet region and is characterized by the significant changes in the flow direction, and (3) wall jet region:
the dominated velocity component is in the radial direction and the boundary layer over the plate is subject to nearly zero pressure gradient and thickens as it moves
radially outwards. Critical review on various aspects of the flow and heat transfer associated with the impinging jets has been conducted by Viskanta [4] and Jambunathan et al. [5]. The impinging jet flow was found to contain a large recirculation vortex around the jet axis and a somewhat smaller adjacent secondary vortex right above the impinging plate in a confined laminar submerged jet (Law and Masliyah [6]). Recently, the recirculating flow resulting from a confined impinging gas jet at low Rej was visualized by Santen et al. [7 & 8], Cheng et al. [9] and Hsieh et al. [10]. It has been noted that the flow of impinging jet can become unstable as the Rayleigh number exceeds certain critical level. Furthermore, Santen et al. [7 & 8]
explained the suppression of the buoyancy induced flow at increasing Reynolds numbers. Hsieh et al. [10] noted that the flow recirculation was in the form of three circular vortex rolls including a primary vortex roll around the jet, a secondary vortex roll in the middle region and a buoyancy-induced vortex roll in the outer zone. The inner and middle vortex rolls are driven by the viscous shear due to a nonuniform velocity distribution in the jet and are stronger and bigger at a high Rej. Hence they are called the inertia-driven rolls. But the buoyancy driven outer vortex roll is important at high buoyancy-to-inertia ratio. The secondary inertia-driven vortex roll only appears at certain high Rej and it is much smaller and weaker than the primary inertia-driven vortex roll. Cheng et al. [9] indicated that increasing the chamber pressure and the temperature difference between the heated plate and air jet caused the outer roll to become larger and the inner roll to become correspondingly smaller.
Moreover, at high buoyancy and inertia the flow becomes time dependent. Hsieh et al.
[10] showed that the vortex flow became time periodic at a certain high buoyancy-to-inertia ratio and the oscillation frequency of the vortex flow increased with Rej. In a rapid chemical vapor deposition (RCVD) chamber, Rayleigh light scattering method was used to measure temperature and visualize the flow by Horton
and Peterson [11]. Their results showed that the flow became unstable at Gr Re2j =5. Recently, heat transfer in confined impinging jets was examined by Hsieh et al. [12].
They concluded that the heat transfer characteristics were only significantly affected by the jet Reynolds number. Moreover, numerical computation using the Reynolds stress model was performed to predict the flow field in confined turbulent jet impingement by Morris et al. [13]. Multiple vortices in the flow were well predicted by the Reynolds stress model (RMS). The k-ε turbulence model was found to be more accurate than the second-moment closure in predicting the turbulent impinging jet flow (Dianat et al. [14]). An experimental study on unconfined impinging jets was conducted to examine flow structure and heat transfer by Carcasci [15] and Angioletti et al. [16]. Their results showed that the convection heat transfer coefficient reached a peak around the stagnation zone for a small nozzle-to-plate distance. More recently, Chung and Luo [17] and Chiriac and Ortega [18] demonstrated that heat transfer rate along the target plate was enhanced by an unsteady impinging jet. As the Reynolds number exceeding certain critical level, a steady to unsteady flow pattern transition for a confined laminar impinging jet with Rej<1000 was numerically investigated by Chiriac and Ortega [18]. They also indicated that the critical jet Reynolds number for the onset of unsteady flow was between 585 and 610. Moreover, the dominant frequency of the unsteady jet flow is in accordance with the primary vortices emanating from the shear layer produced by the jet just issued from the nozzle. The transition between the laminar and turbulent impinging jet flow at Rej=1,500 was suggested by Elison and Webb [19]. A combined experimental and numerical study was carried out by Narayanan et al. [20] to study an impinging slot jet flow. They noted that the secondary peak in the heat transfer coefficient was still high owing to the interaction between the streamwise velocity variance and related motion in the
outer region and to the near-wall turbulence.
Colucci and Viskanta [21] examined the effects of the nozzle geometry on the impinging jet heat transfer. They compared the results measured with two different hyperbolic nozzles and pointed out that the outer peak of the local heat transfer coefficient was dependent upon the geometry of the nozzle. Ashforth-Frost and Jambunathan [22] suggested that the length of potential core in an impinging jet could be affected by the jet confinement and the potential core was longer for a fully developed jet exit velocity profile than a flat jet exit velocity profile. Baydar [23]
experimentally investigated confined impinging jets at low Reynolds number and showed that a low pressure zone appeared on the impinging plate for both single and double jets as the nondimensional nozzle-to-plate spacing (H/D) was less than 2.
Furthermore, heat transfer in multiple jets impinging onto a plate was investigated by San and Lai [24]. They obtained an optimum ratio of the jet-to-jet spacing to jet diameter and proposed a correlation for stagnation Nusselt number. Recently, in our research group Hsieh et al. [25] visualized the detailed flow patterns of a round air jet impinging onto a heated disk confined in a cylindrical chamber and revealed that inclining the chamber top could effectively suppress the buoyancy-induced vortex flow.
In the impinging jet flow encountered in the CVD and RTP processes, the gas jet is at a relatively low flow rate and the buoyancy in the flow is no longer small compared with the jet inertia. Significant flow recirculation can be induced by the buoyancy and the impinging jet flow is driven by the combined effects of the inertia and buoyancy. The importance of the buoyancy on the recirculation flow in a vertical CVD reactor was illustrated by Wahl [26]. Similar investigations have been carried out for various types of CVD reactors including the metal organic CVD reactors [27-30] and single RTP processors [31-33]. In these studies for semi-conductor thin
film deposition [31-33] various vortex flow patterns were reported in the impinging jet flow.