Electromigration Drift Velocity Research Articles

Electromigration in Dual-Damascene CuMn Alloy IC Interconnects

Electromigration reliability of Cu interconnects is an important issue for continued performance improvement of IC technologies. In this paper, we investigate the impact of Mn impurities on the electromigration drift velocity of Cu conductors. Under accelerated test conditions, the drift velocity in CuMn alloys is reduced by an order of magnitude compared with pure Cu with polycrystalline microstructure. Dilute CuMn alloys exhibit a large activation energy for drift of 1.05 eV, which is significantly increased above that for pure Cu as a result of a 0.2 eV impurity(Mn)-vacancy binding energy. There is no measurable impact of the presence of the Mn dopant on diffusion along the Cu/dielectric cap interface in the samples studied here.

Fatal Void Size Comparisons in Via-Below and Via-Above Cu Dual-Damascene Interconnects

AbstractThe median-times-to-failure (t50's) for straight dual-damascene via-terminated copper interconnect structures, tested under the same conditions, depend on whether the vias connect down to underlaying leads (metal 2, M2, or via-below structures) or connect up to overlaying leads (metal 1, M1, or via-above structures). Experimental results for a variety of line lengths, widths, and numbers of vias show higher t50's for M2 structures than for analogous M1 structures. It has been shown that despite this asymmetry in lifetimes, the electromigration drift velocity is the same for these two types of structures, suggesting that fatal void volumes are different in these two cases. A numerical simulation tool based on the Korhonen model has been developed and used to simulate the conditions for void growth and correlate fatal void sizes with lifetimes. These simulations suggest that the average fatal void size for M2 structures is more than twice the size of that of M1 structures. This result supports an earlier suggestion that preferential nucleation at the Cu/Si3N4 interface in both M1 and M2 structures leads to different fatal void sizes, because larger voids are required to span the line thickness in M2 structures while smaller voids below the base of vias can cause failures in M1 structures. However, it is also found that the fatal void sizes corresponding to the shortest-times-to-failure (STTF's) are similar for M1 and M2, suggesting that the voids that lead to the shortest lifetimes occur at or in the vias in both cases, where a void need only span the via to cause failure. Correlation of lifetimes and critical void volumes provides a useful tool for distinguishing failure mechanisms.

Relationship between interfacial adhesion and electromigration in Cu metallization

A relationship between the adhesion of a Cu conductor to its surrounding medium, the electromigration drift velocity, and lifetime in a conventional electromigration test has been demonstrated. Lifetime measurements demonstrate that the drift velocity and the activation energy for mass transport along an interface correlate with the measured adhesion energy of the interface as determined by a four-point bending test. The results indicate that a linear relationship is expected between the electromigration activation energy and the intrinsic work of adhesion, which is consistent with a simple model relating the two. The data indicate that interfacial cleanliness and bond character at the interface (i.e., metal/dielectric versus metal/metal bonding) have a significant impact on all measured parameters.

Numerical Simulation of Grain-Boundary Grooving by Level Set Method

A numerical investigation of grain-boundary grooving by means of a level set method is carried out. An idealized polycrystalline interconnect which consists of grains separated by parallel grain boundaries aligned normal to the average orientation of the surface is considered. Initially, the surface diffusion is the only physical mechanism assumed. The surface diffusion is driven by surface-curvature gradients, while a fixed surface slope and zero atomic flux are assumed at the groove root. The corresponding mathematical system is an initial boundary value problem for a two-dimensional equation of Hamilton–Jacobi type. The results obtained are in good agreement with both Mullins analytical “small-slope” solution of the linearized problem (W. W. Mullins, 1957, j. Appl. Phys. 28, 333) (for the case of an isolated grain boundary) and with the solution for a periodic array of grain boundaries (S. A. Hackney, 1988, Scripta Metall. 22, 1731). Incorporation of an electric field changes the problem to one of electromigration. Preliminary results of electromigration drift velocity simulations in copper lines are presented and discussed.

Level set modeling of transient electromigration grooving

A numerical investigation of grain-boundary (GB) grooving by means of the level set (LS) method is carried out. GB grooving is emerging as a key element of electromigration (EM) drift in polycrystalline microelectronic (ME) interconnects, as evidenced by a number of recent studies. The purpose of the present study is to provide an efficient numerical simulation, allowing a parametric study of the effect of key physical parameters (GB and surface diffusivities, grain size, current density, etc.) on the EM drift velocity as well as on the morphology of the affected regions. An idealized polycrystalline interconnect which consists of grains separated by parallel GBs aligned normal to the average orientation of interconnect's surface is considered. Surface and GB diffusions are the only diffusion mechanisms assumed. The diffusion is driven by surface curvature gradients and by an externally applied electric field. The corresponding mathematical system is an initial boundary value problem for a two-dimensional Hamilton–Jacobi type equation. To solve the electrostatic problem at a given time step, a full model based on the solution of Laplace's equation for the electric potential is employed. The resulting set of linear algebraic equations (from the finite difference discretization of the equation) is solved with an effective multigrid iterative procedure. The details of transient slit and ridge formation processes are presented and compared with theoretical predictions on steady-state grooving [E. Glickman, M. Nathan, J. Appl. Phys. 80 (1996) 3782; L. Klinger, E. Glickman, V. Fradkov, W. Mullins, C. Bauer, J. Appl. Phys. 78(6) (1995) 3833; L. Klinger, X. Chu, W. Mullins, C. Bauer, J. Appl. Phys. 80(12) (1996) 6670]

Electromigration drift velocity in Cu interconnects modeled with the level set method

Electromigration (EM) drift velocity (DV) experiments in polycrystalline pure Cu lines are simulated numerically with the level set method. The simulation is based on a grain boundary (GB) grooving model, incorporating an electric field. The model is distinguished by two key requirements imposed at the triple point where two surfaces and a GB meet: that of GB and surface flux coupling (flux continuity), and that of permanent equilibrium between surface and GB tensions. Surface diffusion exists only at the advancing cathode edge, and is driven both by local curvature gradients and by the local field. Using independent, literature diffusivity values, the simulation yields both the DV prefactor and the EM activation energy in an Arrhenius-type expression. An excellent match is obtained with experimental DV values in the T range of 573–723 K. Some implications regarding the material transport mechanism are discussed.

Electromigration Modeling of Blech Experiment with Comparison to Recent Experimental Data

AbstractFinite element analysis is applied to simulate electromigration in a copper line/tungsten junction, which is a typical structure in Blech experiments adopted to measure the electromigration drift velocity by edge displacement. Because the specific resistivity of tungsten is higher than that of copper, the electric current crowds at the corner where the underlying tungsten bar connects to the edge of the copper line. Therefore, material in the corner is depleted faster than that in other places of the copper/tungsten interface. Our simulation reveals that a stress that may be high enough to crack the interface is also built up in the corner. The evolution of the line resistance under various current densities is calculated from our model and compared to recent experiment results.

Electromigration Drift Velocity in Al‐Alloy and Cu‐Alloy Lines

Electromigration has been investigated using both drift velocity and resistance techniques in pure Al, Al(Cu), pure Cu, Cu(Mg), Cu(Zr), and Cu(Sn) isolated lines overlapping W studs (vias between metallization levels), or lines overlaying a W underlayer line. The phenomenon of void growth was investigated in sections of lines which extended past the cathode terminal to form reservoirs through which no current flowed. This has been modeled using the electromigration‐induced vacancy wind. The activation energy for electromigration was found to be 0.60 eV and 0.9 to 1.1 eV for multigrained and bamboo‐grained pure Al structures, respectively, and 0.77 eV for multigrained pure Cu thin film lines. The drift velocity of Cu alloys has been investigated using Mg as the solute, which enhances the Cu drift velocity, while Zr or Sn drastically reduce the Cu migration rate.

Electromigration and Diffusion in Pure Cu And Cu(Sn) Alloys

AbstractAtom movements of Cu in pure Cu and Cu(0.5 to 2 wt.% Sn) alloys have been investigated using drift velocity and radioactive tracer techniques. The void growth rate in pure Cu at the cathode end, as a result of electromigration driving force, linearly increases with the applied current density. A marked decrease in the Cu grain boundary diffusivity and electromigration drift velocity is attributed to Sn trapping of Cu atoms and/or binding defects at grain boundaries. The effect is more pronounced at lower temperatures. The activation energies for electromigration and diffusion of Cu in pure Cu grain boundaries were found to be in the range of 0.7 – 0.9 eV. Addition of about 0.5 to 2 wt.% Sn increased these energies to 1.1 – 1.3 eV, respectively which resulted in enhancement of electromigration resistance by several orders of magnitude over that in pure Cu at the field operation temperature.

Electromigration and Diffusion in Pure Cu and Cu(Sn) Alloys

AbstractAtom movements of Cu in pure Cu and Cu(0.5 to 2 wt.% Sn) alloys have been investigated using drift velocity and radioactive tracer techniques. The void growth rate in pure Cu at the cathode end, as a result of electromigration driving force, linearly increases with the applied current density. A marked decrease in the Cu grain boundary diffusivity and electromigration drift velocity is attributed to Sn trapping of Cu atoms and/or binding defects at grain boundaries. The effect is more pronounced at lower temperatures. The activation energies for electromigration and diffusion of Cu in pure Cu grain boundaries were found to be in the range of 0.7 - 0.9 eV. Addition of about 0.5 to 2 wt.% Sn increased these energies to 1.1 - 1.3 eV, respectively which resulted in enhancement of electromigration resistance by several orders of magnitude over that in pure Cu at the field operation temperature.

Grain Boundary Grooving as the Mechanism of the Blech Electromigration in Interconnects

AbstractWe present a new approach to understand the mechanism of "homogeneous", or Blech electromigration (EM). This phenomenon describes macroscopically homogeneous displacement of the up-wind edge of thin film lines in microelectronic devices and is responsible for openings at contact windows, "vias" and other sites of perfect diffusion flux divergence.Our SEM, EPMA and EM drift velocity experiments have revealed the gradual transition from the microscopically homogeneous EM displacement to the highly nonhomogeneous mode wherein copious islands of residual material remain behind the drifting cathode edge of aluminum stripes. The transition is shown to occur due to an increase in either the current density, j, or in the stripe length, 1. The latter case suggests, that the transition results from the growth of the net grain boundary (GB) diffusion flux, I=le-Ib ,where Ie∝j and 1b∝1/1 are the EM flux and stress-gradient-driven back flux, respectively.Based upon recent progress in the theory of GB grooving under "external" GB fluxes, with surface diffusion acting as the healing mechanism, grooves' propagation along the line and their merging is considered to be the micromechanism of the "homogeneous" EM. In terms of the simple model described, the transition from the slow receding of the cathode butt edge slightly wrinkled by shallow grooves (A-regime of EM) to the fast extension and merging of slot -like grooves (B-regime) accounts for the transition observed in EM mode, while in both regimes the EM displacement velocity, V, is presumed to represent the groove propagation rate.The theory developed reduces to Blech formulae for V for the truly homogeneous A-regime and predicts quite different EM kinetics for the B-regime of microscopically nonhomogeneous EM. The latter is expected to dominate for films loaded by high current density with large grains and low surface diffusion.The dependence obtained for the residual mass left behind the drifted edge vs the displacement velocity, V, for unpassivated aluminum stripes of various lengths, loaded by j=2-106 A/cm2 at 548K provides a good evidence in support of a new approach.

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 in the two-level interconnect yield an activation energy of 0.58 eV for Al, in contrast to 0.78 and 0.83 eV for Al(0.5 wt % Cu) and Al(2 wt % Cu), respectively. The activation energies of the electromigration drift velocity were found to be 0.86 and 0.68 eV for Cu and Al in Al(2 wt % Cu, 3 wt % Si), respectively. These results enable one to infer that the kinetic process is controlled by electromigration of Cu along grain boundaries instead of by dissolution of the Al2Cu precipitates.

Electromigration in Cu/W Structure

Mass transport by electromigration in sputtered Cu line segments on a continuous W line has been measured using the drift velocity technique at temperatures from 166 to 396 °C. The Ta/Cu/Ta line segments are patterned by dry etching techniques. Cu mass depletion (voids) at the cathode end and accumulation (hillocks) at the anode were measured as a function of time from scanning electron microscope micrographs. The edge displacement of Cu was found to increase linearly with time. The activation energy for Cu electromigration drift velocity, which relates to the product of effective charge number and diffusivity, Z*D, is found to be 0.6 eV.

Eiectromigration in Cu/W Structure

ABSTRACTMass transport by electromigration in sputtered Cu line segments on a continuous W line has been measured using the drift velocity technique at temperatures from 166 to 396 °C. The Ta/Cu/Ta line segments are patterned by dry etching techniques. Cu mass depletion (voids) at the cathode end and accumulation (hillocks) at the anode were measured as a function of time from scanning electron microscope micrographs. The edge displacement of Cu was found to increase linearly with time. The activation energy for Cu electromigration drift velocity, which relates to the product of effective charge number and diffusivity, Z*D, is found to be 0.6 eV.

Electromigration in Al/W and Al(Cu)/W Interconnect Structures

ABSTRACTThe electromigration drift velocity of Al in Al(3wt.% Si), Al(2wt.%Cu), and Al(2wt.%Cu,3wt.%Si) was measured in a temperature range 133 to 220 °C with current densities of 1.0 to 1.5×106A/cm2. In Al(3wt.% Si), a significant Al depletion at the cathode end and accumulation at the anode end of stripe were observed within a few hours at 1.5×106A/cm2 and 200°C. In addition, local hillocks and voids along the metal lines were observed. For Al(Cu,Si), the Al drift velocity was slowed down by Cu addition. The majority of hillocks started to grow at a distance about 6 μm away from the cathode end with current density of 1.5×106 A/cm2. The drift velocity of Al in Al(Cu,Si) was found to be a function of time starting with an initial low value and increasing to a an final steady-state value. The behavior was attributed to the migration of Cu and dissolution of Al2Cu precipitates. The activation energies of the depletion 3 Aμm of Al(2%,Cu, 3%Si) was determined to be 0.90±02 eV. The dissolution and growth of A12Cu in the tested samples of Ti/Al(2%Cu)/Ti/TiN were observed using the scanning electron microscope and an electron microprobe.

The effect of anodization on the electromigration drift velocity in aluminum films

The electromigration drift velocity has been measured, using the Blech–Kinsbron [Thin Solid Films 25, 327 (1975)] edge displacement method, for aluminum thin films under different thicknesses of anodization. The drift velocity is found to decrease with increasing anodization thickness. The results are interpreted in terms of a change in the self-diffusivity of aluminum as a result of the compressive stresses imposed by the anodized layer. The effect of the variation of diffusivity with stress on the stress distribution within a drifting thin-film sample is discussed.

Reduced aluminium electromigration in future integrated circuits — A problem of test procedure and threshold mechanisms

In aluminium interconnect lines of integrated circuits there are different electromigration failure modes such as stripe interruptions and contact failures. For the conditions demonstrated here, the electromigration drift velocity limits the lifetime of metallization by contact failure long before interruptions would occur. In short aluminium lines, as well as in all Al-Cu lines, the mass flow is reduced by a grain boundary electromigration threshold. Considering different threshold mechanisms, grain boundary and interface electromigration can be eliminated in large-grained narrow lines. The remaining, drastically reduced bulk electromigration will not be detectable under operating conditions.

Lifetime and drift velocity analysis for electromigration in sputtered Al films, multilayers, and alloys

Besides high lifetimes and a low standard deviation of failure times, the electromigration drift velocity is an important parameter for metallization reliability. Here, lifetimes should be high and drift velocity low. For pure Al films, only properly chosen and optimized sputter conditions yield good film quality and electromigration data comparable to evaporated stripes. Annealing and oxide covering improve lifetime and reduce drift velocity of pure Al structures. For redundant AlTiAl multilayers, annealing yielded an enormous lifetime enhancement of 3 orders of magnitude. However, these stripes are not suited for practical applications because of whisker formation and of an increased drift velocity! These disadvantages do not appear for a lifetime-improving TiN intermediate layer instead of Ti. A further method of lifetime enhancement is Cu-addition which reduces the drift of stripes. Finally, for (Al + Cu)TiN(Al + Cu)-structures, lifetime is improved even further in combination with low drift velocity.

The threshold current density and incubation time to electromigration in gold films

The electromigration in thin gold films was studied by means of direct drift velocity measurements. A threshold in the electromigration drift velocity was found, and its value is inversely proportional to the gold stripe length. Electron-beam-deposited films, in contrast to sputtered ones, also demonstrated the presence of an incubation time for the cathode displacement. The electromigration activation energy was found to be 0.67–0.72 eV and the effective charge -0.3 to -186. In addition, alloyed Au-10 vol.% Mo specimens were examined. These films showed a much slower mass transport from the anode to the cathode—the opposite direction to that of non-alloyed gold films.

Electromigration in thin aluminum films on titanium nitride

The aluminum electromigration drift velocity was measured at the temperature range 250–400 °C. A threshold current density was found inversely proportional to the stripe length. An activation energy of 0.65 eV was found for the drift velocity. The occurrence of the threshold is explained by opposing chemical gradients created by the atom pile-up and depletion at the stripe ends. The threshold may explain several observations reported previously. The threshold is increased by decreasing the temperature or by enclosing the aluminum in silicon nitride. Virtually no electromigration is seen for very short aluminum stripes even at current densities above 106 A/cm2.