M. B. Small

Electromigration in Al(Cu) two‐level structures: Effect of Cu and kinetics of damage formation

The electromigration characteristics and kinetics of damage formation for Al(Cu,Si) line segments on a continuous W line and Al(Cu)/W two‐level interconnect structures have been investigated. The mass transport as a function of temperature was measured using a drift‐velocity technique. The flux divergence at the line/stud contact was found to be responsible for formation of open failure in the interconnect structure, as shown by a direct correlation observed between mass depletion at the contact and resistance increase of the line/stud chain. The depletion of Al at the stud contact is preceded by an incubation period during which Cu is swept out a threshold distance from the cathode of the line. This leads to a damage formation process which is controlled by both Cu electromigration along grain boundaries and dissolution of the Al2Cu precipitates. This is distinctly different from single‐level interconnects measured using a conventional electromigration test site. Measurements of the mean failure lifetime...

Electromigration in two-level interconnect structures with Al alloy lines and W studs

It is demonstrated that electromigration testing needs to be performed in structures that reflect use conditions, such as when there is a flux divergence as provided by the W stud‐Al(Cu) interface rather than in a simple planar structure. The Al(Cu)/W interface has been investigated using both drift velocity and resistometric techniques with pure Al, Al(0.5 wt. % Cu) and Al(2 wt. % Cu) lines on W studs for interlevel connections. It is shown that the mass depletion can be correlated to the resistance change and electromigration failure in line/stud chains. A new effect is demonstrated in that a critical length of Al has to be depleted on Cu before the Al can migrate; when such migration starts the Al catches up with the Cu rich region, leading to slower motion and the production of extrusions which will also cause failures by shorting to adjacent lines.

Electromigration failure due to interfacial diffusion in fine Al alloy lines

Damage formation at grain boundary junctions has long been recognized as the dominant electromigration failure mechanism in metal lines. We report the results of drift‐velocity experiments on fine lines with no reservoirs and find that the interfacial mass transport, along the edges of the lines, is faster than that along grain boundaries. This causes mass depletion at the cathode end of the line, leading to electromigration failure. The result demonstrates a new failure mechanism due to electromigration in submicron lines with bamboo grain structures.

A phenomenological study of meniscus lines on the surfaces of GaAs layers grown by LPE

Abstract The phenomenon of “meniscus lines” on the surfaces of liquid phase epitaxial layers of GaAs is described, and it is shown that these are associated with a “stick-slip” motion of the edge of the liquid as it is removed from the surface. Mechanisms which may account for the creation of this phenomenon are postulated.

Growth and dissolution of ternary alloys of III–V compounds by liquid phase epitaxy and the formation of heterostructures

Abstract In the LPE process, in contrast to other epitaxial processes, solid and fluid phases are close to equilibrium at their interface. Thus the underlying material (either the substrate or the last formed layer) can interact with the solution and hence can actively participate in the formation of the current layer. This is particularly true of multicomponent III–V systems, where each successive layer is formed from the same elements, albeit in different proportions. We review a model based on a minimum set of assumptions, we solve the appropriate transport equations, and we discuss this model in the light of recent experiments. Our conclusions include: (i) The computation of diffusion profiles is closely coupled to that of interface motion. (ii) When the solid tends to dissolve, the heteroepitaxial layer forms by solid state diffusion of material exchanged between liquid and solid through the moving interface. (iii) Compositional profiles in the solid depend crucially on departures from an exactly saturated condition in the liquid. (ic) For small enough solid diffusivities and liquid undercoolings, the LPE system behaves as if it were in a state close to equilibrium. (v) Diffusion coefficients in the solid are larger during dissolution than during growth. This is probably due to defect injection. (vi) There is no reason, experimental or theoretical, to believe that the average growth rate can ever change sign during a processing step, unless the system is subjected to temperature programming.

The formation of Ga1−xAlxAs layers on the surface of GaAs during continual dissolution into Ga‐Al‐As solutions

When solid GaAs is placed in an undersaturated solution of Ga, Al, and As, it has been observed that a layer of the solid Ga1−xAlxAs forms on the surface. In the past the presence of this layer has been attributed to a process of regrowth following sufficient dissolution to saturate the solution. On the other hand, an analysis of the kinetics of the situation has suggested that dissolution should be continuous and that the surface layer is formed by solid diffusion. An experiment is reported here in which the solid is forced to dissolve continuously, and a layer of Ga1−xAlxAs of similar thickness to those reported by others has been found to be produced on the surface. Such a layer could not have been produced by regrowth. In order to be produced by solid diffusion, the diffusion coefficient of Al in GaAs must be anomalously high.

Growth and dissolution kinetics of ternary III‐V heterostructures formed by liquid phase epitaxy. III. Effects of temperature programming

The kinetics of growth and dissolution and compositional grading in the solid are computed when a liquid‐phase epitaxial system is subjected to temperature programming. We solve the transport equations (including the effect of interface motion) by means of series expansions in powers of t1/2. This requires a careful analysis of the phase relations that hold at the crystal‐fluid interface, because these are now explicitly time dependent. The theory is applied to a specific active layer of a (GaAl)As heterostructure laser, but the calculations can be extended to any III‐V ternary system, and indeed to any multicomponent system.

An explanation for the phenomenon of meniscus lines on the surfaces of (GaAl) as alloys grown by LPE

Abstract Fresh evidence is presented that meniscus lines are formed at the intermediate surface of a multilayered structure grown by LPE and that both departing and arriving solutions generate such lines. The shape of a large meniscus line has also been determined and compared with the theoretically predicted shape for a moving groove associated with a mobile grain boundary. The correspondence between the two shapes and the similarities between the physical situations lead to the conclusion that the two phenomena are essentially the same except that a moving solution - gas phase boundary replaces a mobile grain boundary. This suggests that the phenomenon is fundamental and possibly also relevant to other crystal growth situations.

Anomalous diffusion behavior of Mg in GaAs

This experiment was undertaken to demonstrate that the anomalous diffusion behavior reported of Al in the alloy system (AlGa)As is not unique. The interdiffusion of Al and Ga across the buried heterojunctions between the AlAs and GaAs layers of a superlattice structure has been shown to be slow, whereas very much faster diffusion has been observed into the free surface of a GaAs wafer exposed to a saturated solution containing these two elements and Al. In this letter it is shown that Mg exhibits similar behavior and, further, a consequence is that when Mg is used to dope GaAs layers grown by liquid phase epitaxy the dopant penetrates to a significant depth in the underlying solid.

Conditions for constant growth rate by LPE from a cooling, static solution

It is important to be able to grow crystalline layers at a constant rate in order to have good control over thickness, morphology, and composition. If LPE layers are grown from thin solutions over a relatively small temperature range an essentially constant growth rate can be achieved. The necessary conditions for constant growth rate are: (i) that t/gt⪅0.1, where t is the growth period, and τ the relaxation time associated with an Arrhenius liquidus curve and a steady cooling rate b; τ is given by τ = mC0b, where m is the slope of the liquidus line at the initial temperature and C0 is the intial concentration of the solution, and (ii) that t/τ⪆a2. The parameter a is in the range 3 x 10−2 to 10−1 where a is defined by the equation a=l(DmC0b)12, in which l is the effective linear extent of the solution and D the diffusion coefficient of growth units in it.