The principle of spray deposition (SD) was firstly proposed by A.Singer at the University of Swansea in the UK during the 1970s [1]. This process can prepare near net-shape products with less oxide contamination and fine grained, equiaxed and cellular final structure [2-4]. The spray deposited preform is not fully dense, usually containing 10-15 per cent porosity, dependent on the thermal and solidification conditions of the liquid droplets during deposition [5]. The tensile strength of the preform, which depends on the density and metallurgical bond of its constituent particles, is very low, and the elongation is virtually zero.

In order to prepare a fully dense metal strip by the spray deposition route, a green preform is first prepared by spray deposition, followed by forging, extrusion or rolling using the appropriate processing techniques to achieve the required mechanical properties as well as eliminating porosity. Understanding the densification, deformation and fracture behaviour of porous metals during the forming operation is of great importance.

Densification of porous metal preforms by hot working requires shearing the pores so that they collapse and the porosity is eliminated [6]. Unlike the densification of porous metal by close-die forge or extrusion via a high hydrostatic stress field, densification of porous metal during rolling occurs under a generalised stress field having both hydrostatic and deviatoric components.

The application of such a stress field on porous materials causes flattening of pores and brings about rapid densification by collapsing the pores at a faster rate. Rolling porous materials is an effective fabrication process for obtaining a fully-density product from a spray deposited porous preform. However, elongation and spreading in longitudinal and transverse directions is a disadvantage - forming surface and side cracks during rolling when the materials lack sound metallurgical bonds.

An optimal design of such a process requires a proper understanding of the effect of different variables on the deformation behaviour of porous materials. In hot densification rolling of porous metal strip, densification is achieved essentially by the longitudinal flow occurring in the strip that, in return, is related to the thickness deformation given to the strip and its instantaneous relative density. In the case of hot rolling of a fully dense material, since the constancy of volume is maintained, a decrease in thickness results in the increase in length and spreading in width. However, in the case of the porous materials, volume changes take place during rolling and the longitudinal and transverse flow is significantly lower than that produced in a fully dense material when given the same amount of thickness deformation.

Although theoretical investigations for understanding the rolling behaviour of fully dense materials have been carried out, there are comparatively few studies of rolling porous metals [7, 8]. Shima and Yamada [7] predicted the density and thickness of strips obtained by the roll compaction of loose metal powder using the upper-bound approach. The plane-strain hot rolling of sintered porous plates was modeled by Mori and Osakada, using the finite-element method [8]. However, a considerable gap still exists in the understanding of the effect of rolling parameters on deformation and fracture behaviours.

In the present work, an attempt has been made to analyse the strains in deformation zone and the effect of rolling parameters (packing materials, original width of workpiece and thickness reduction) on deformation behaviours (the spreading, elongation and changes in volume of porous workpiece) and fracture behaviors. A novel rolling processing named Frame-Pack rolling is developed which can improve the workability of porous workpiece effectively during rolling.

The spray deposited material used in this investigation was 4032, as given in Table 1, with a relative density of about 90 per cent. The relative density (r) is defined as the ratio of apparent density to theoretical density. The spray deposition process parameters are listed in Table 2.

Test 4032 preforms, 26mm in thickness, 50mm in width and 90 mm in length, were used to investigate the deformation behaviour (spreading, elongation and reduction in thickness) in the rolling deformation zone of porous metal. A schematic representation of the rolling deformation zone is shown in Figure 1. A test preform was packed with stainless steel on both surfaces. The hot rolling was performed on a two-high mill with rolls of 360mm in diameter and 600mm in length. The nominal rolling speed was 0.43m/s. The workpiece was preheated for approximately one hour in a furnace operated at a fixed temperature about 733ºK. Thickness was reduced by 30 per cent in one pass.

More samples were rolled with different packing materials. Materials with three different friction factors were used. They were aluminum, stainless steel and stainless steel with a rough surface - artificially made nicks on the surface of stainless steel gave a higher friction factor. Rolling comprised preheating (approximately one hour) and reheating (approximately 30 minutes) between each pass in a furnace operated at a fixed temperature about 733ºK. Packed preforms were hot rolled for three to five passes in which the nominal reduction per pass was approximately 10 per cent.

A 4032 test preform, 16mm in thickness, 49.1 mm in width and 88.4mm in length was used to test frame-packed rolling. To investigate the effects of the frame on the workability and fracture of a porous workpiece, the 4032 preform was placed in a frame made from porous 8009 alloy with stainless steel on top and bottom surfaces. A schematic representation of the frame pack technique is shown in Figure 2. The relative density of 8009 alloy is 80~85 per cent which was less than that of the 4032 preform. The size of the frame was 16mm in thickness, 49.5 mm in width and 88.8mm in length. The framed preform was hot rolled for three passes in which the nominal reduction per pass was approximately 10 per cent.

Fractional thickness deformation, fractional longitudinal elongation, fractional lateral spreading and fractional changes in volume were calculated from the dimensions of samples before and after hot rolling. An average of four to five measurements was taken for each parameter.

The results of this work are divided into three sections: deformation characteristics of porous metal workpieces in the deformation zone; effects of packing materials and original width on deformation behaviours of porous workpieces, effects of frame on the deformation and fracture behaviours.

The deformation characteristics of porous metal workpieces in the deformation zone are presented in Figure 3. From Figure 3(a)-(c), it is observed that the transverse strain near to the surface is lower than that near to the centre, while the thickness strain is characterised by an opposite trend. The longitudinal strain behaviour near to the surface and centre of the deformation zone were found to be two curves crossed at neutral point which occurs closer to the exit. In the backward slip zone, the longitudinal strain near to the surface is higher than that of near to the centre. In the forward slip zone, the longitudinal strain near to the surface was found to be lower than that of near to the centre. This is explained easily in terms of the friction between roller and workpiece that hinders the metal flow. The effect of friction is decreased from the surface to the centre. However, friction accelerates longitudinal flow of metal in backward zone while blocking it in the forward zone.

Figure 4 shows the effects of thickness reduction on strain variation across the thickness of porous workpieces. It is clear that up to the thickness reduction of about 40 per cent, the strain variation across the workpiece thickness is symmetrical around the centre of the workpiece, and the strain is higher in near-surface areas relative to the centre of the workpiece, which is different from the fully dense material. As for fully dense materials, when the thickness reduction is small, compressive deformation mainly takes place near to the surface. Strains across the workpiece thickness are uneven. Strain hardening in metals near the surface makes deformation in the centre possible with the increase of thickness reduction, which results in a relatively even strain across the workpiece thickness. However, as regards porous materials, the presence of pores in the matrix will result a lower work hardening effect during deformation relative to fully dense materials. Therefore the compressive deformation still occurs mainly in the metals near to the surface even where the thickness reduction is up to about 30 per cent. Friction between roller and workpiece makes it reduction difficult near the surface of fully dense metal, but porous metal can be compacted more easily near the surface due to the pores.

The effects of packing materials on the spreading, elongation and changes in volume of porous workpieces are shown in Figure 5. Different packing materials give different friction condition at the roll/workpiece interface. From the graphs shown in Figure 5, the spreading and elongation are observed to decrease with the friction factor, while the reduction in volume is increased with it, i.e. the densification rate is accelerated with the friction factor.

In materials processing, a workpiece of simple shape is deformed plastically into a final complex form, hopefully without defects. If the workpiece contains voids initially, the task of processing is considerably more difficult. Since voids form as an intermediate step in the development of ductile fracture in fully dense materials, their pre-existence at the initiation of plastic deformation seriously limits the amount of metal flow that can be achieved before fracture [10]. The presence of highly dispersed porosity makes the preform crack easily.

It is noted that the performs are liable to cracking or tearing when the thickness reduction is up to a critical value, other than at the initial stage during rolling. This may be because that during the initial stage of rolling, rolling serves mainly to compact the strip without effecting much extension, and overall stresses acting on the strip, compressive in thickness and tensile in the rolling direction, are too small to cause any plastic deformation of the material [11]. It may be assumed that any longitudinal stress component tending to produce cracking is small. However, though metal particles are not plastically deformed, it is possible that stress concentration near pores may cross the yield point of the material thereby causing the sliding of the particles in the direction of rolling and restacking themselves in void regions available in their vicinity. Transitional stacking of particles results in densification. So most of the compressive strain given to the preform in the thickness direction is consumed without plastically deforming the material. Very little elongation in the rolling direction observed at early stage of rolling could thus be understood.

The efficiency of densification is higher in the initial stage of rolling. With increasing thickness reduction, after attaining a relative density of about 0.95, the density and compressive strength of material increases rapidly. The plastic deformation becomes the predominant densification mechanism, the spreading and elongation in rolling rapidly increase and some of the flow goes into elongation without further closing the pores. The efficiency of densification becomes low and the material still lacks a sound metallurgical structure. Once the tensile stress in the rolling direction exceeds the fracture stress, microcracks on the surface perpendicular to the rolling direction occur. In addition, the reduction of surface temperature of the workpiece during rolling due to the contact of roller and workpiece results in decreased working ability and the appearance of microcracks. Two approaches are effective in improving the workability of porous materials during rolling - keeping a uniform temperature across the thickness of the workpiece and restraining longitudinal and transverse metal flow during early stages to speed up densification and obtain sound metallurgical bonds.

It is clear from Table 3 that fracture strains increase with friction, which can be induced by different packing materials. For difficult-to-roll fully dense materials, hot pack rolling, usually packing the workpiece in a ductile and sacrificial material (cans), is an effective method to improve the rolling ability. These cans minimise heat losses, as well as oxidation or other contamination of the workpiece during hot rolling. As for porous materials, they play another important role increasing the friction between roller and workpiece, providing high friction on the flow of metal during rolling and more effective densification.

In this experiment, two packing materials are used: aluminium (soft can, whose yield stress is lower than workpiece) and stainless steel (hard can, whose yield stress is higher than workpiece). The results show that the fracture strain of aluminium-packed workpiece is lower than that of the stainless steel-packed. This trend may be because that the softer aluminium layer undergoes a larger elongation during rolling, resulting in a tensile stress on the surface of workpiece surface which is liable to give birth to the cracks on surface. However, a hard can provides an inverse effect on the flow of the workpiece during rolling and the restraining effect is intensified with a rough surface.

In order to impose a more intense restrain on the elongation and spreading on the metal during rolling, another packing process has been developed, i.e. the Frame-Pack method. There is a gap only of about 0.4mm between the frame and workpiece to allow the workpiece into the frame.

The mechanism of a porous frame in restraining the flow of the workpiece both in transverse and longitudinal direction is shown in Figure 6. During rolling, the reduction in thickness of frame is the same as that of the workpiece. The spreading and elongation of frame and workpiece are shown in Figure 6 designated Db1, DL1 and Db2, DL2, respectively. It is clear that Db1 acts as an restrain to Db2. the elongation of frame (DL1) will be less than that of the workpiece (DL2) when given to the same thickness reduction due to a relatively low density.

The frame imposes an effective restraint on the elongation of the workpiece, while at the same time, the frame around the sides proves to be of benefit in maintaining a uniform temperature in the workpiece during the rolling process. From the results of the investigations reported in this paper, the following conclusions may be drawn.

(1) Friction, caused by different pack materials, affects the deformation behaviour by a decrease of lateral and longitudinal flow, resulting in accelerated densification.

(2) Lateral flow of metal decreases with the original width while the densification speed increases.

(3) A porous frame around the sides of porous workpiece can improve the workability effectively by imposing restraints to the lateral and longitudinal flow that results in an effective densification.

The team
This feature was abstracted from Deformation and fracture behaviour of a porous 4032 alloy prepared by spray deposition during hot rolling, prepared by Chen Zhen-hua, Zhan Mei-yan and Xia Wei-jun.

Acknowledgements
The authors would like to thank Chen Gang and Teng Jie for the preparation of porous preforms. Thanks are also due to Li Jian-jian and Wu Heng for their careful preparation of rolling workpieces.

References
1. Singer A R E. Met. Matter., 4 (1970) 246
2. U Prakash, T Raghu, A A Gokhale, S V Kamat: J. Mater. Sci., 34(1999)5061
3. E J Lavernia and N J Grant: Mater.Sci.Eng., 98(1988)381
4. D J Skinner, R L Bye, D Raybould and A M Brown: Scipta Metall., 20(1986)867
5. S Hariprasad, S M L Sastry, K L Jerina et al. : Metall.Trans., 24 A (1993) 865
6. S Bhargava, R K Dube. Metall.Trans., 19A(1988)1205
7. S Shima, M Yamada. powder. Metal., 27 (1984):39
8. K Mori and K Osakada. Inter Mech Sci 29(1987)229
9. CHEN Zhen-hua, in Multi-layer spray deposition technology & applications (Hunan University publishing, Chinese, 2003) p.192
10. Howard A Kuhn. applied mechanics division, AMD, 16(1976)171
11. R K Dube. Inter. Mater. Rev., 35 (1990) 253

Print this Article Return to Features Index