Home » KnowledgeCenter » How to: Cold Bump Pull (CBP) » How to: Cold Bump Pull (CBP)
This extensive how-to consists of 13 paragraphs and 7 pages. Please use the navigation on the right.
This manual is a guide advising what to consider and how to best setup a Cold Bump Pull (CBP) test. It generally refers to solder ball testing but the principles apply equally to most bump testing. Hence, some say Cold Ball Pull testing.
CBP stands for Cold Bump Pull or Cold Ball Pull. Our CBP test type page contains a video of 80µm cold bump pull.
To grip a ball we have to reform it to the shape something like a mushroom. This is done with special tweezer jaws that have cavities that reform the ball when they close. You might think reforming is bad but think about a shear test. In a shear test the reforming at a typical test force is often more and a lot of it over the bond.
If we get a higher test force by “over gripping” the ball so that the reforming cuts over the bond this is better than a lower force with a ductile or extrusion where the bond was still not tested but at a lower force.
Ideally the “stem” of the mushroom shape is the larger or the same diameter as the bond but if over gripping tests the bond at a higher force, it is a better test. We might still get a ductile or extrusion when we over grip, but the bond has been tested to the highest force possible.
Not all balls can be effectively gripped by the CBP test method. The ball height to diameter affects what can be tested. The ratio of what can be tested is roughly defined by diameter / height < 2,2.
As is always the case xyztec’s Golden Rules of Bond Testing apply;
For more information on the Golden Rules refer to Science of bond testing module BTM1.1.
In the case of CBP testing we are interested in the strength of the bump to the pad and/or the pad to the substrate. This is the “bond”.
The construction of the bond varies depending on the application. There can be a few or many layers in the bond but in any case we want to know how it fails and at what force.
We have already decided that Pull Testing is more like the true loading condition. That is to say it is not shear.
If the bond is stronger than the bump we can get one of the following failure modes: ductile or extrusion. In this case we want to setup the test to get the highest force possible.
Note, if the force of extrusion is only slightly higher than a ductile failure it may be better to set the test to get ductile failures. This is because extrusions can contaminate the jaws with debris much more rapidly, making cleaning more difficult. See also: paragraph 11: Cleaning the jaws.
To recap: we always want the failure mode of interest and if that is not possible, we want the highest force.
To reform and grip a ball with the maximum force possible the jaw quality must be maintained. It is important to monitor jaw quality during use.
The two jaws must also align as good as possible. The maximum error of any misalignment side to side and in closing height is 5% of the ball diameter. Cavity contamination is also a cause of poor jaw quality that can reduce the grip force. See the section on jaw cleaning.
The first stage of doing a test is the alignment of the jaws to the bump. This comprises of the jaw opening and then centering the jaws above the bump.
The jaw opening must be set such that they do not hit the bump or others around it when they descend to grip. If the bumps are very close to each other you have to either shear them away or use jaws with back relief.
Set the test method to “return to the start position”.
Before starting a test the jaws must be aligned over the bump. First set the Z height so the bottom of the jaws is just above the tops of the bumps. The closer they are the easier it is to make the alignment in X and Y, but be sure you have sufficient clearance not to touch the highest bump. As you have set the method to return to the start position this height will remain when you align for the next test.
Once you have set the height, align over the bump so the cavities are concentric with it.
Warning: do this next step very carefully!
To get used to the X and Y alignment and to set an initial jaw closing force you need to do a test close on a bump. Choose a bump and align over it. Then manually slowly drive down to just touch the substrate.
Now by observing the bump look at your XY alignment. Try reforming another bump and again look at your XY alignment. Repeat until you are happy with the result and know how to align the jaws.
The jaws take time to close and the bump takes time to reform. Bumps of less than 100µm reform almost instantly, where as a large bump of 750µm can take up to 5 seconds. This is why when we were testing the first grips we waited for 5 seconds.
When you are testing your grip and alignment observe how long it takes for the jaws to stop moving. In the test method program there is a safe closing time delay (“Holdtime”) to ensure the bumps will be fully reformed at the closing force selected.
When a test is started the Z axis drives the jaws down to touch the substrate and then to a programmable Touchdown force. This force depends on the application. It should be as low as possible but sufficient to ensure the jaws come into good contact with the substrate. When the bumps are very close to each other the landing force can help the jaws slide past the adjacent bumps. This is fine providing the bumps are not significantly damaged.
Typically “Touchdown” forces to get you started should be:
Best results are obtained by trial and error but initially set the landing force as shown above.
Most CBP tests require the jaws to close in contact with the substrate. This gives the strongest grip. In the case of metal defined pads a ”distance from surface” may be required to avoid damaging the bond when the bump is gripped and reformed.
Metal defined bumps often have low height and require a ”distance from surface” resulting in a poor grip force. On the other hand the bond is not supported by the mask and so shear testing is often a better option.
Do a few tests with your initial settings, recording the failure modes and test forces.
Most likely you will get extrusions but if you get bond failures lower the grip force until you do get an extrusion. Then increase the grip force a little until you just get all bond failures.
If you get extrusions, increase the closing force until you get either a bond failure or a ductile failure.
If you cannot get a bond failure, try different closing forces to find the one that gives the highest test force. The failure mode can then be extrusion or ductile. For balls above 100µm it will probably be a ductile failure mode. For balls below 100µm it is sometimes an extrusion.
If you are increasing the closing force, watch how long it takes for the bump to reform because you may need to increase the “Hold time”.
You can now review all of the settings. It may seem complicated but the simple objective is to get bond failures (failure modes of interest) or the highest test force.
Warning: If you are increasing the closing force observe how the test force changes. If the test force starts to go down, there is no point increasing the closing force any further and you can break the jaws if you carry on increasing it.
Test speed will have very little effect in standard CBP. This is because the jaws are locked onto the sample by the grip and the axis has only the landing force distance to acceleration, so all tests are relatively slow. For bumps up to 300µm a test speed of 100µm/s is recommended. Above 300µm you can use 200µm/s to increase the number of tests per minute. At these speeds the axis can accelerate to the programmed speed in the landing force distance.
Xyztec’s USB tweezers have programmable control of all the features needed for CBP. These are:
There are many ways the test can be set up but a typical test sequence is;
The construction of solder balls varies a lot. Variables include;
There are also many different bump alloys, pad alloys, pad surface finishes, reflow processes and aging. Because of this variety there are no industrial standards for bond strength. As a general guide based on ductile failure and the strength of solder the maximum test force will be in the region shown in this graph.
The cavities of the jaws will become contaminated by debris from the bump. To maintain good and consistent testing this must be cleaned away. It can be done manually, but be very careful not to damage the jaws and their cavities. Solder can be automatically cleaned off using xyztec’s patented contactless cleaning system (click here to view a video). Contact your local representative for details.
The following five failure mode definitions, that can be further divided into eight types, have been set by JEDEC. In this how-to we already mentioned most of them. We refer to the JEDEC website for full details.
The standards JEDEC JESD22-B115, JEITA EIAJ ET-7407 and IPC-9708 apply to Cold Bump Pull testing.
The determination of failure modes is also called grading. As vision technology progresses, it is expected that machines will be able to grade failure modes even more effectively than humans. With its advanced Halcon vision library and high-resolution perpendicular cameras with smart lighting systems, several customers already use the Sigma to accurately determine failure modes in solder bumps.
Local support and parts, global presence, world-class testers
J.F. Kennedylaan 14b5981 XC PanningenThe Netherlands
Schäferei 1806237 LeunaOT GünthersdorfGermany
72/7 M.12 Soi. SoonthornwipakBangpla, Bangphli,10540 Samut PrakanThailand
No. 157, Zhongzheng 6th St., Hukou Township, Hsinchu County 303, Taiwan (R.O.C.)
Room 2012 Haichuang Mansion, No.288 Dengyun Road, High-tech district, Kunshan, Jiang Su, China
Follow us on:
© 2021 xyztec bv
Xyztec develops world-class bond testing technologies and works together with global partners to provide local support worldwide.