How the double lap joint tops the peel strength and T₂60 methods.

For some time, the accepted test method for measuring adhesion in the PCB industry has been the peel strength test. Yet it is accepted that this technique has been less than adequate in guarding against delamination failures during reflow and wave soldering. A recently introduced test method, now a part of IPC-TM-650, is the so-called "T₂60 Method" in which a thermal event is imposed that causes a delamination of the test specimen. The objective of this study is to compare the stress field associated with a board delamination to that generated by the peel and T₂60 tests.

Initial attention is given to a first principals characterization of the stress field associated with a uniform, free expansion of a PCB, as in the reflow process. If severe enough, this will result in a delamination. The stress fields produced by the peel and T₂60 tests are then analyzed and compared to that of a free expansion. Finally, a novel technique is suggested for performing this measurement.

According to the first principals of materials, the fundamental stress field of a differential cube at equilibrium is uniquely defined by three shear stresses acting along each of the faces of the cube and three normal stresses acting normal to each of the faces. All other reactions can be expressed in terms of these six variables. Any other reaction is redundant. For a 2D analysis, as is the case here, the fundamental stress field reduces to two normal and two shear stresses (Figure 1). When these four stresses are known, the stress field is uniquely defined regardless of the stimulator (e.g., thermal, mechanical, etc).

FIGURE 1. The fundamental stress field showing two normal and two shear stresses, per the first principals of materials.

Unfortunately, peel strength alone does not uniquely define the stresses shown in Figure 1 and, consequently, the measurement is not completely definitive. That is, the peel strength along with some other variables (not yet discussed) must be defined in order to reduce the measurement to the fundamental variables of Figure 1. As most in the industry are aware, the peel strength causing a rupture may vary if these additional variables are not the same. Later, these additional variables will be identified.

Stress field at delamination. The failure mechanism of interest is a rupture causing a breach within the PCB. Most often, this occurs between a copper layer or feature and a layer of prepreg. When this failure is observed, it is normally the result of a thermal event such as reflow or wave soldering. The rupture is caused by the dissimilar lateral expansion between the copper and the adjacent glass-epoxy layer. This creates a shear stress between the two layers which, if severe enough, will cause a rupture; i.e., a delamination.

The analysis considers a structure composed of two dissimilar components bonded together (Figure 2). The structure is unrestrained and initially at a temperature T₁. The temperature is elevated to T₂. An equilibrium stress analysis of this phenomenon is given in Reference 2.

FIGURE 2. Basic structure of two dissimilar components bonded together.

After temperature T₂ is achieved the stress field can be described by the free body diagram in Figure 3.

FIGURE 3. Free body diagram of structure after temperature is elevated.

Rayleigh's Law requires
ΔL = αL(T₂-T₁)

and according to Hook's Law
F/A =(ΔL/L)E

where α is the coefficient of thermal expansion (CTE)
E is the modulus of elasticity
L is the length of the sample.

Consequently
ΔL_A= Δ_A L (T₂-T₁)+ π_x L² / E_At_A
ΔL_B= Δ_B L (T₂-T₁)- π_x L²/ E_Bt_B
ΔL_A = ΔL_B
t is the thickness of the component, A or B.

Therefore
π_x= ΔT(α_B- α_A) E_A (t_A /L)/[1+(t_AE_A/t_BE_B)]       (1)

Since there are no normal forces applied, both normal stresses (σ) are zero. It is important to realize that π_x is the only stress generated in an unrestrained thermal expansion of two joined dissimilar materials. If the strength of the bond between component A and B is less than π_x, rupture occurs.

It is interesting to note scale factors t_A /L (geometry) and t_AE_A/t_BE_B (structural) appear in the equation, which is consistent with the observation that thicker copper is more susceptible to delamination. Obviously, for a laboratory simulation of the delamination of an actual board using a scaled coupon, the test specimen must reflect these two parameters as they occur in practice.

Peel strength. The preferred technique for measuring the robustness of an adhesive joint has for many years been the peel strength test. Several methods are available for this measurement in IPC-TM-650. The test is usually performed on an Instron. The test coupon is composed of a copper foil that is laminated to a PCB substrate. A small width of laminated copper foil is pulled vertically from the sample and the force required is measured. This force divided by width of the copper foil is referred to as the peel strength. Often this test is performed at an elevated temperature in an attempt to account for the thermal effects of assembly.

Attention is now given to the fundamental stress field developed in the peel strength measurement (Figure 4). The free body diagram is in Figure 5.

FIGURE 4. The fundamental stress field that occurs during peel strength testing.

FIGURE 5. The free body diagram of the peel strength measurement.

Summing the forces orthogonal to the radius in Figure 5 gives

And by definition

The tensile stress in the copper at the point that the force F is applied is

Then after integrating and applying the boundary condition

It follows that the shear stress applied by the adhesive (π_a) is

Summing forces in the radial direction shows that

Or

It is interesting to note that the ratio of the two stresses is

where
θ is the polar position on the copper strip (measured in radians) G is the shear modulus of elasticity of the adhesive F is the applied force, e.g. the peel force E_c is tensile modulus of elasticity for copper r is the radius of curvature of the copper strip t is the thickness of the copper.

To an order of magnitude the ratio of

And

or more

The order of magnitude of the ratio of the shear stress to the normal stress is, therefore, unity and both stresses are of consequence in the peel strength measurement.

The stress vector resulting from these two orthogonal stresses then causes a fracture in the plane defined by the stress vector. Also, the radius of curvature, which generally is uncontrolled, is seen to play a strong role in the peel strength measurement.

In a thermal event, the fracture is parallel to the copper or glass/epoxy interface and only the shear component is present. The peel strength measurement is then at best problematic from two aspects; it interjects a superfluous tensile stress as well as incorrectly identifying the plane of failure. To say the least, the measurement is suspect.

T260 test method. A second index that is coming to the forefront for measuring bond strength is referred to as the T260 test method. In this case, a sample of an actual board is placed into a thermal mechanical analyzer (TMA). The TMA chamber is quickly heated to 260°C and held. A very sensitive probe is placed on the top of the sample to detect any vertical expansion of the test sample. When a sudden expansion is detected, as caused by a delamination, it is sensed by the TMA and the elapsed time recorded. The elapsed time is then the index of structural integrity of the sample. This technique has the advantage of faithfully simulating the stress field associated with a thermal event.

The time required to generate a failure in the T260 test procedure is often in excess of 20 min. and consequently some elect to accelerate the test by using a higher equilibrium temperature, in some cases up to 280°C. The risk, as always, is the higher the acceleration the greater the likelihood of triggering additional failure mechanisms. IPC has also defined an abbreviated form of the T260 test. In this case, the coupon is floated on a solder pot at 260°C for a specified time. This stress is repeated until the coupon fails and the total time of exposure recorded.

Unfortunately, the end point metric (time) is an indirect measurement of the shear strength. The major disadvantage of the T260 test is that the failure occurs at a temperature somewhere in between ambient and the temperature of the test chamber. Most likely, there are uncontrolled thermal gradients in the test sample that enhance the shear stress. There is no convenient way to relate the endpoint measurement, time-to-failure, to the shear stress. Although this procedure has some obvious advantages over the peel strength measure, it is far from perfect.

Double lap joint. A proposed option to overcome the disadvantages of the previous techniques is the double lap joint (Figure 6). As shown, a double lap joint is formed by laminating copper between two pieces of laminate. The copper is segmented and overhangs at the ends of the specimen. (It is essential that the rear vertical surfaces of the copper not be bonded to the laminate, as this will add an undesirable element of structure.) The ends of the overhang are placed in jaws of an Instron and pulled until failure occurs. As indicated by the free body, the shear stress at the copper laminate interface is
π = (F/2)/(Xd)

FIGURE 6. The formation of a double lap joint.

where d is the thickness of the sample.

The stress is independent of the thickness of the copper and the laminate. To ensure a fracture along the copper laminate boundary, the thickness of the copper and laminate should be large. After failure, the sample should be inspected to ensure that the failure occurred along the copper/laminate interface. If not, the sample should be redesigned. This test can be conducted at an elevated temperature by using a heated chamber as is often used in Instron testing to measure material properties at high temperatures.

The analysis shows that the stress field generated by the peel test is poor reflection of that caused by a free expansion such as in reflow. The T260 test generates the desired stress field, but the measurement matrix is time-to-failure, which is an indirect index. A third option, a double lap joint, is shown to overcome all of these issues, but the design and fabrication of the test sample is critical.   PCD&M

REFERENCES

1. Patrick Brooks, Hugh Roberts and Kuldip Johal, Oxide Alternative Qualification, A Case Study, Atotech Corp., 2003
2. B. E. Gatewood, Thermal Stresses, McGraw-Hill, 1957
3. J. L. Parker and Patrick Brooks "A Comparison of PCB Adhesion test Methods and Adhesion Promoters," IPC Printed Circuits Expo, February, 2005

Dr. J. Lee Parker is an independent consultant with JLP Consultants. He has a Ph.D. in aeronautical engineering and holds nine patents. He can be reached at This email address is being protected from spambots. You need JavaScript enabled to view it..