INVESTIGATION OF THE THERMAL DEFORMATION OF ELECTRONIC
PACKAGES WITH ELECTRONIC SPECKLE PATTERN INTERFEROMETRY
Hans Reinhard Schubach
Dantec Dynamics GmbH
Kässbohrer Str. 18, 89077 Ulm
ABSTRACT
Thermal expansion has a great influence on the durability of electronic components. Joining of different materials exposed to
variations of temperature generates the strain mismatch which can be locally strong. These strains might cause cracks which
initiate malfunctions.
On the other side simulations (FEM, …) are used in the electronic industry, but the material parameters (coefficient of thermal
expansion, young’s modulus, …) have to been known. A validation of results is necessary. In some cases simulations are not
possible, than the measurement precise deformation is necessary.
In this paper we will show the investigation of thermal deformation of electronic packages with electronic speckle pattern
interferometry (ESPI).
Introduction
Electronic components are used in the automotive, communication, aerospace and other industries. Miniaturization, higher
package density and accelerated development processes have a great impact on the reliability of electronic components.
Rapid changes of ambient temperature or internal production of heat may occur during operation. This may create high
thermal stresses due to the mismatch of the thermal expansion coefficients of the different materials in electronic components.
For example of the thermal expansion coefficients for typical materials are shown in the table below.
•
Gold:
α= 14,2 *10-6/K
•
Solder(95 Sn-5Ag):
α = 23,2 *10-6/K
•
Silicon:
α = 2,6 *10-6/K
•
FR4:
α = 13,0 *10-6/K
As you can see the variation of the thermal expansion coefficient can be up to a factor of 25.
The 3D-ESPI measuring system Q300 is a very powerful tool to characterize the reaction of electronic components on thermal
loads. Laser Speckle Interferometry is a technique to measure full field and contact less the 3D deformation of objects under
investigation. The Field of view can vary from square meter down to square millimeter.
3D ESPI MEASUREMENT PRINCIPLE
Electronic speckle pattern interferometry (ESPI) [1], [2], [3] is a full field optical deformation measuring technique, which is
know for the high spatial as well for the high displacement resolution. This technique has been applied to standard material
testing [4],[5]. Recent developments have extended the capability of the measuring technique to the thermal characterization
of MEMS and electronic components.
During the measurement an object under investigation is illuminated by coherent laser light. For in-plane measurements the
sample is illuminated by two object beams, for out-of-plane measurements the illumination is done by one object beam and
one reference beam. The reflected laser light is detected by a CCD camera. Due to displacements on the sample the
interference pattern changes on the CCD. The speckle interferometric principle enables to measure full field deformation. The
combination of various directions of illumination and one direction of observation enables the determination of full field three
dimensional deformations.
Figure 1. Principle of 3D Measurement with speckle interferometry
By adaptation of the magnification by the appropriate choice of lenses and the sensitivity by long illumination arms a standard
3D ESPI System can be used for the measurement of small object with high sensitivity. A field of view down to 1 mm² is
possible. Components with large dimensions can be inspected with focus on the area of interest which can be of small
dimensions. The spatial resolution is depend of the field of view and is better 1/1000 of the diagonal. The sensibility of the
displacement measurement is in the range 0,05 µm for the in-plane and 0,01 µm for the out-of-plane measurement
DEFORMATION OF YAW SENSOR DUE TO POWER CONSUMPTION
The measurement set-up is shown in figure 2. The optical sensor head of the Q-300 is equipped with a macro lens and long
illumination arms for optimizing measurement sensibility. The surface was inspected applying a power supply voltage of 5 V
and monitoring the surface deformation during heating up. The power consumption was 1,35 mV. In figure 4 the resulting
displacement fields are shown. The maximum displacement in the in-plane direction is in the range of 0,6 µm and in the
range of 0,3 µm for the out-of plane displacement. The z-displacement field after 10 minutes shows the place of the sensor
element based on silicon (lower part) and the ASIC (upper part)
Figure 2. Set-up of 3D-ESPI system
Figure 3. Sensor in a PLC44 housing
6.138
6.138
After 40 sec
4.680
0.095
0.017
0.009
0.166
3.262
0.127
1.762
0.089
3.262
0.001
1.762
-0.008
3.262
1.762
0.063
0.344
0.050
0.344
-0.016
0.344
0.031
-1.114
0.011
-1.114
-0.025
-1.114
-0.001
-2.532
-0.027
-2.532
-0.033
-2.532
-3.991
-0.066
-0.033
-0.065
-0.105
-5.409
-0.097
-0.050
-5.409
4.680
0.146
0.54
0.41
1.762
0.12
0.344
-0.02
-0.15
-0.42
0.195
4.680
3.262
0.25
-0.29
6.138
6.138
0.66
4.680
-5.409
-7.327 -5.898 -4.429 -3.000 -1.490 -0.061 1.408 2.837 4.347 5.776 7.245
-7.327 -5.898 -4.429 -3.000 -1.490 -0.061 1.408 2.837 4.347 5.776 7.245
6.138
0.52
-3.991
-0.058
-0.143
0.65
0.39
-0.041
-3.991
-7.327 -5.898 -4.429 -3.000 -1.490 -0.061 1.408 2.837 4.347 5.776 7.245
After 10 min
4.680
4.680
0.159
0.127
6.138
0.205
0.191
3.262
0.098
3.262
1.762
1.762
0.049
0.16
0.344
0.000
0.344
-1.114
0.04
-1.114
-0.049
-1.114
-2.532
-0.09
-2.532
-0.097
-2.532
-3.991
-0.21
-5.409
-0.34
0.29
-0.55
-0.146
-3.991
-0.195
-5.409
-3.991
-5.409
-0.243
-0.46
-7.327 -5.898 -4.429 -3.000 -1.490 -0.061 1.408 2.837 4.347 5.776 7.245
-7.327 -5.898 -4.429 -3.000 -1.490 -0.061 1.408 2.837 4.347 5.776 7.245
x direction
y direction
-7.327 -5.898 -4.429 -3.000 -1.490 -0.061 1.408 2.837 4.347 5.776 7.245
z direction
Figure 4: Displacement fields after switching on the power supply voltage of 5V (Power= 1,35mW
Thermal expansion of the substrate FR4
The heat loading of a FR4 sample has been done in temperature heating unit. The temperature change was from room
temperature up to 145°C. The deformation fields have been measured every 5°C. The set-up is shown in figure 5. Figure 6d
shows the FR4 sample as seen be the sensor. The size of the sample was 20 mm x 20 mm. The resulting displacement fields
are visible in figure 6a to 6c. The spatial distribution of CTE’s is shown in the same figure 6e and 6f. The thermal expansion in
the horizontal (x-) direction is quite uniform, which is reflected in the homogenous CTE (see figure 6e). The CTE is calculated
as thermal strain divided by the temperature change. The deformation in the vertical direction in not homogenous, the CTE in
vertical direction is changing strongly (see figure 6f). The change is periodic with the period of the upper uniaxial fiber layer.
ESPI Sensor
Temperature
controll
unit
Temperature
Chamber
Figure 5: Measurement set-up for the measurement of FR4
A
b
c
D
E
f
Figure 6: a: x-displacement field in µm; b: y-displacement field in µm; c: z-displacement field in µm; d: live image of the FR4
Sample as seen by the sensor; e: CTE in x-direction [10-6 1/K]; f: CTE in y-direction [10-6 1/K]
Measurement of Ball Grid Array (BGA)
Figure 8 shows the scheme of the design of the assembly of a BGA. There are several components with different CTE’s. The
assembled BGA undergoes a reflow process which means that the whole assembly is heated up to 240°C and cooled down.
The measurement setup is shown in figure 5. The component is just placed in the heating chamber whose temperature is
controlled by the electronic control device. Fig 9 shows the sample as seen by the sensor. The thermal expansion of BGA is
measured from room temperate up to 140 °C.
Figure 8: Schematic design of a BGA
Figure 10 shows the in-plane and out-of-plane displacement fields after a thermal loading from room temperature up to 140
°C. The mismatch in the thermal expansion can be seen in figure 10A and B. The higher thermal expansion of the FR4 results
in a large bending of the whole assembly (figure 10C). Figure 11 shows the variation of the in-plane and out-of-plane
displacement vectors along the dashed line indicated in fig 10.
Figure 9:The BGA assembly as seen by the sensor.
A
B
C
Figure 10 In-plane and out-of-plane displacement fields of a BGA assembly after thermal loading. A: inplane displacement in xdirection, B: inplane displacement in y-direction, C: bending (out-of-plane displacement); all displacement values in µm.
out-of-plane displacement [µm]
in-plane displacement [µm]
35
30
25
20
15
10
FR4
BGA
5
0
-5
12
10
8
6
4
2
0
-2
-4
-6
-8
-15
-15
-10
-5
0
5
10
-10
-5
0
5
10
x-coordinate [mm]
x-coordinate [mm]
A
B
Figure 11: A: variation of the in-plane displacement vectors along the dashed line indicated in fig 10A, B: variation of the outof-plane displacement vectors along the dashed line indicated in fig 10C:
Non destructive control of Flip Chip packages for space applications
Shrinking dimensions of flip chip assemblies (figure 12) make inspection of bumps or solder joints always more difficult as
standard non destructive control techniques reach their resolution limits. For the tests two groups of samples were prepared.
The first one is composed of assemblies with all their 130 bumps. The second group consists of flip chip assemblies with
defects. In order to evaluate the 3D-ESPI system sensitivity, "large" defaults were intentionally introduced and half of the
bumps (the inner ones) were removed.. With the Q300 3D-ESPI measurements were performed between room temperature
and 125°C.
Figure 12: Flip chip assembly ( GaAs die connected on AlN substrate) for 3D ESPI measurements. Flip chip samples were
manufactured with 4x5.5mm², 100 µm thin, GaAs MMICs connected to AlN substrates
Figure 13a represents the measured out of plane displacement (or z-deformation) experienced by a chip with all its bumps :
0.5 micron z-deformation is observed at the center of the die The observed strain can be attributed only to the small coefficient
of thermal expansion (CTE) mismatch between the GaAs die (CTE ~ 6 ppm/°C) and the AlN substrate (CTE ~ 4 ppm/°C .
Figure 13b exhibits the results of the same measurement performed on a die where half of the bumps were removed. Here the
z-deformation at the center of the die is higher than 3 microns! The large difference between these two measurements is
summarized on Figure 14 with the illustration of the out of plane displacement of both packages along their diagonal.
(a)
(b)
Figure 13: (a) 3D-ESPI out of plane displacement measured between 25°C and 125°C on the assembly from figure 12. The
die is connected with all its bumps. (b) Same measurement for a die with only half of its bumps. The grey value scale
indicating the z-displacement is the same on both figures.
Figure 14: Out of plane displacement in the diagonal of flip chip assemblies with (red line) and without (blue line) defects.
Conclusions
Measurements with the Dantec Dynamics laser optical measuring system Q-300 enable us to characterize thermal loading not
only of electronic components. Thermal stresses can be easily detected, defects will be found. The sensitivity and spatial
resolution is quite well adapted to the application. Due to the variability of the sensor, it can be easy adapted to the
measurement problem. The validation of FEM calculations is possible (and done).
References
1.
2.
3.
4.
5.
Angel F Doval, Meas. Sci. Technol. 11 (2000), R1ff
Pramod K. Rastogi (Editor); Optical Measurement Techniques and Applications; Artech House 1997, ISBN 0-89006-516-0
R.S. Sirohi, Speckle Metrology, Marcel Dekker, New York, 1993.
A. Ettemeyer; Recent Developments of Speckle Interferometry Technologies for Industrial Applications; ASNT Conference
Austin, 08.-12.03.2004
H. R. Schubach; Investigation on Laser Welded High Strength Steel with 3D-Fullfield Optical Methods; Test 2003,
Nurnberg; 13-15.5.2003