K&S-4700 Convertible Dual Wire Bonder

Wire bonders literally bridge the gap from nano- and micro-structured patterns¹⁾ to millimeter-sized devices²⁾, by allowing the user to weld a conductive wire between the two 'domains'. At its core, wire bonding is a form of solid-state welding—a process in which two materials are joined without melting, relying instead on pressure, friction, and sometimes heat to form a metallurgical bond.

The K&S-4700 Convertible Dual Wire Bonder provides wedge-wedge and ball-ball bonding with the same machine. The system is equipped with a Nikon SMZ660 optical microscope (0.8-5x zoom). At the moment, Al wire is available for bonding.

Additional information:

• Link to the manual.
• For questions and requests, contact Luc Wigbout.

In wedge bonding, a fine wire is pressed onto a bond pad using a wedge-shaped bonding tool while subjected to ultrasonic vibration, resulting in a solid-phase weld between the wire and the substrate.

The bonding tool vibrates laterally (parallel to the substrate) at ultrasonic frequencies—typically in the range of 60–120 kHz. This vibration is driven by a piezoelectric transducer in the bonding head. When the wire is clamped between the tool and the bond pad, the ultrasonic motion causes friction at the interface.

This friction scrubs away surface oxides and contaminants, enabling metal-to-metal contact. The localized frictional energy also induces plastic deformation of the wire and slightly heats the bonding interface, both of which help activate atomic diffusion and interfacial adhesion. This results in a solid-state bond, without melting either material.

To prevent resonance effects, make sure that the adhesion between the sample and the chip carrier is good.³⁾

The wedge tool applies downward force while vibrating, causing the wire to deform plastically into the surface of the bond pad. The amount of deformation depends on bonding parameters (ultrasonic power, time, force, time) and material properties (wire hardness, pad roughness). Optimal deformation results in a “stitch bond” with a well-defined footprint and strong mechanical/electrical connection. Too little deformation can cause weak bonds, while excessive force can damage fragile substrates or tear the wire.

Repeated thermal cycling (e.g. during transport measurements) can stress the bonds mechanically, so bond quality and wire routing (avoiding strain) are critical.

When the wire is fed through the wedge and also through the hole at the backside of the wedge correctly (which requires some time or handyness), wedge bonding can start. First make sure that the surplus of wire within a reasonable amount or do a test bond first.

Wedge bonding is a uni-directional process, starting the first bond at the front of the sample and making the second bond at the back of the sample (while sitting in front of the machine).


A detailed view of the wirebonder wedge is shown below.

• Ultrasonic Power (P) (typical range: 1-2):
  The ultrasonic power applied determines the amount of energy applied to the bond, by varying the amplitude of the ultrasonic vibrations.
  
  Higher power increases interfacial friction and bond strength but can also damage soft metals or delicate structures.
  
  If the bond still doesn't set at powers > 2, you should check if your sample is mounted correctly. For instance, if the sample is resonating with the needle while it's applying the ultrasounds, there is less friction between, and therefore, no bond will be set. Increasing the power will not solve that.
  
  If you keep trying to bond at high powers, there is a chance that you will set a bond between on the wedge itself, resulting in a clogged wedge. This means the wedge needs to be replaced, which means downtime.

• Bond Time (T) (typical range: 1-5):
  Duration the tool is in contact with the wire while vibrating.
  
  Longer time allows more energy to transfer, but can overheat or deform the bond.

• Bond Force (F) (typical range: 1-2):
  The downward pressure from the tool. The bonding force is applied to the wire, while the ultrasonic power is being applied. It consists of a static force which is determined by counterweights in the machine itself and an amount applied by the electromagnetic coil. The latter one is determined by the force dial. The value indicated by the dial is merely used to make a recipe which works for specific applications.
  
  Too little force causes poor bonding; too much can crush pads, crack capping/passivation layers, or thin the wire excessively.

Finding the right parameters for you process can be a tedious job and requires some work and experience. Please check the logbook if there are any similar samples so that you can use those parameters (P/T/F) as a starting point. Be aware that if you go out of their typical ranges, you are probably not on the right track…

• Bond doesn't work. There is nothing on the sample.
  There is not enough friction: increase Force or Time.

• Bond doesn't work. There is a pancake of wire on my sample.
  There is way too much friction. Decrease Time or Power.

• Bonds are weak, they break easily.
  You might be flattening the wire too much: decrease Force!

• Bonds are set, but when moving to the loop height, they still break.
  Lower the Loop Height to about the same height as the 1st Search Height. Upon setting the first bond, slowly increase the Loop Height. If the bond still breaks when you do this very carefully, you might want to perform a reverse stepback-manoeuvre to make sure that the tension is not on the bond, but in a kink in the wire. You can find out how to do this here.

• None of these work?
  If changing the Time and Force doesn't work, try changing the Power. If the Power > 2, and it still is not working, the problem lies in your sample, not the wire bonder. Make sure that the sample is mounted rigidly, so that no resonance effects occur. Make sure that the sample is flat, so that the foot of the wedge is optimally touching the sample surface.

• Search height:
  The search height determines the height at which the bond head stops before making the actual bond. This allows for fine positioning of the wedge over the bondpad. The search height should be 75-100 μm above the bondpad surface (~3 times the wire thickness). A different search height can be specified for the first and second bond when there is a height difference between the first and second bondpads. When the first and second bond are in the same plane, both search heights should be equal. (If you want to bond in manual mode it is still important to set the correct search height, this should be done in semi-auto mode.)
  
  Make sure that the search height is above the substrate surface! This prevents the needle from crashing into your sample, potentially destroying it. If you want to play it save, increase the search height to the maximum before putting your sample underneath the needle.

• Loop height:
  After the first bond is made, the wedge is set to the loop height, which is dependent on the bond length needed, whether or not the first and second bond are in the same plane and so on. The value set is an absolute value and does not depend on the search height. This means that when the search height is changed, it is probable that the loop height needs to be adjusted as well.
  
  If you decide to increase the loop height, more wire will be used for your bond. If you accidentally have too much wire, and you want to decrease the loop height, you will see that the wire does not go back in the bonding tool. Use the loop height wisely.

• Tail length (typical value: 5):
  For wedge bonding, the tail length determines the wire length which is produced after the second bond. Thus, it is also the amount of wire with which the first bond starts.
  
  Because the K&S-4700 is a relatively old system (>20 yrs), this tail is not very reliable anymore. Between every set wire bond, you should check the tail using the microscope! If the tail is not enough, feed the wire manually.
  
  (For ball bonding, it also determines the wire length after the second bond, but this length then is a critical factor in the Electric Flame Off (EFO) process as the amount of wire determines the size of the ball. This means that too little tail produces a too small ball (which can get stuck in the capillary) and too much tail will result in problems when doing the EFO.)

• Kink height & reverse (optional)
  Normally when the wedge moves to loop height, the first bond has to withstand extra stress because the wire is pulled from it. This can make the bond weaker or even loosen it completely. To avoid this, a kink can be made just after the first bond, before making the loop. This is done by adjusting the kink height and reverse dials.
  
  The first of the two determines to height to which the wedge will rise (and thus the wire length protruding the bond). Adjusting the reverse dial will influence the amount of reverse motion of the table, after the the wedge has risen to kink height. After the table has reversed, the wedge will rise to loop height. This way, when rising to loop height, the kink will be under stress, not the first bond.

• Stepback (optional)
  The stepback determines the distance the motorized table moves between the first and second bond pads (if they are at a regular distance this is useful). The maximum stepback setting is 6.4 μm.
  To use the stepback the motor has to be turned on at the right button panel of the machine.
  
  For a short animated video that shows the kink height/reverse and stepback so that a nice bow bond is made, click here.

• Check if the wedge is in the microscope's field of view at minimum magnification. Focus on the wedge tip.
• Check if a small tail of wire is visible. If not, feed manually.
• Without touching the spool of the wire, check that the wire is not too tense. If the wire looks straight, carefully loosen the nut on the left. This makes sure that the spool can unwind.⁴⁾

• Check if the wedge is installed correctly. With the test button on the left panel you can check if the wedge is installed properly in the US transducer. Push the button down, the 'U/S' LED should light up continuously. If bonding doesn't work make sure you check the US transducer again. If not, adjust the wedge height.

• Check if the clamp is closed.
• Check if the 'Tear' knob is at 5.
• Check if the 'Tail' knob is at 5.
• Check if 'Semi-Automatic' bonding is selected (right side switch)

• Mount your workpiece⁵⁾ onto the workholder, positioning it in the middle of the workholder. Then push the clamps up and slide them a bit inward so that the workpiece is fixed by the clamps when they are released.
• Determine how you will bond the sample; where will the first bond be, and where will the second one go
• First check that the wedge is correctly mounted: the flat side of the wedge is facing the user and the height is set correctly. The height can be calibrated with the wedge install tool.
• Check that the wire has been fed through the metal clip, black clamp and wedge correctly.

• Set Loop Height: 2
• Remove sample holder, so the bonds are made in air
• Press and release left mouse button: this makes a bond in air
• Now the wedge is at loop height. Adjust work holder such that it fits under the wedge.
• While looking through the microscope, adjust the work holder height such that the wire is just above the lowest point of the sample
• Set loop height to a high value so you can easily remove the work holder and make the second bond in air
• Set 1st and 2nd Search Height to the same values.
• Place work holder back and press and hold left mouse button.

With everything setup, you can now start bonding in semi-automatic mode.

Wedge bonding is a uni-directional process, starting the first bond at the front of the sample and making the second bond at the back of the sample (while sitting in front of the machine). When the wire is fed through the wedge and also through the hole at the backside of the wedge correctly (which requires some time or handyness), wedge bonding can start. First make sure that the surplus of wire within a reasonable amount or do a test bond first.

1. Press and hold the left mouse button
2. At search height, position the wedge over your first bondpad
3. release the left mouse button. Do not make any sudden movements with the mouse while in the middle of the bonding process.
4. Adjust the loop height until the wedge is high enough above the sample surface to suit your needs.
5. Check wether the second search height is set correctly for the second bond
6. Move to the second bondpad and press and hold the left mouse button
7. Fine position over the second bondpad.
8. Depress the left mouse button to finish the second bond.

Always leave the system with the wire in the wedge, write the logbook, clean your workspace and turn of the goose neck lightguide if you used it.

• Open clamp (flip switch on the right)
• Move clamp lifter up (clamp lifter is on the left of the machine, close to the arm)
• Close clamp
• Move clamp lifter down

If you can see that the wire is not going through the lower hole, but instead, is curling up, this could mean that the wedge is clogged. here you see how to fix that.

A nice bond needs to have a nice bow to avoid tension and stress. If a wire is connected straight between contact pads there will be tension on the bond pads. When making the second bond you want to overshoot over the position you want to bond, then push back the wire to the bonding pad, this way you're releasing tension and create a nice bow in your bond.

Ultrasonic energy is transferred via the transducer to the wedge. It is important that the wedge is installed properly and energy can be transferred without problems. It is important to check this before and during bonding.
On the left panel use the 'TEST' switch, the U/S LED between the 1st and 2nd should light up continuously. \\When loosening or tightening the screw of the wedge always make sure to use the correct height gauge! The height setting of the wedge is extremely sensitive.

If the wedge is clogged, you might want to clean it. Find a small petridish and fill it with a bit of isopropanol. Hold the petridish underneath the wedge without spilling, and use the Test-switch to turn ultrasounds on. This way, we basically ultrasonicate the wedge as if it were in a ultrasonicator. Sometimes, 10 seconds is enough to clean the wedge. If it is still clogged, you might want to remove it. Please contact a technician before you do that.