Material Considerations in Ultrasonic Welding
by Tom Kirkland
Several years ago, a panicked business owner called asking for help figuring out why his ultrasonic welding operation had stopped working. I had talked with him on the phone a few times, but had never seen his operation. I committed half a day that week to troubleshooting for him.
Arriving at his plant, I looked over his parts. They seemed to fit well, and it was the sort of application that should work – an electronics assembly with two black plastic shells joined at the midline to produce something that looked a lot like a large squirt gun.
I asked him the first question you learn in engineering school about problem-solving: "If this worked last week and doesn't work this week, what has changed?"
His answer was that nothing had changed, but – oh, by the way – his molder was using a new, less-expensive material. The old material: ABS. The new material: 30-percent talc-filled polypropylene.
Well, there’s the answer.
The business owner, however, was insistent that the new material would work... because his molder had told him so... so, I set the welder to weld for about 15 seconds and cranked up the clamp force.
"This won't hurt your machine or your tooling," I said, "but, if this has any possibility of welding with the new material, it should at least tack it a bit."
I hit the buttons, and after about eight or nine seconds, smoke rolled out from under the horn. When the horn retracted, the upper part was stuck to it, dripping a little molten material onto the electronics and the other shell, which were still sitting in the fixture. The energy director was pounded flat a bit, but there was no noticeable tacking anywhere.
"Oh, no," the owner said, sounding like I had just spilled mustard on the upholstery of his new Cadillac. I assured him it hadn’t hurt the welder or tooling and explained that he simply needed to go back to the other material. However, the solution wasn’t so simple.
"I just took delivery of ten thousand sets," he said, "and it took a month to get them."
The moral of the story? Know thy materials.
Plastic materials 101
The two main families of plastic materials are thermosets and thermoplastics. Thermosets are formed into shape and baked in the mold to produce crosslinking, in which branches of adjacent polymer molecules grow together to form what essentially is one giant molecule. Applying another heat cycle will not cause flow and, if temperature becomes high enough, the material simply will burn without flowing. For this reason, welding of thermosets is impossible.
Thermoplastic molecules retain independence and do not crosslink on application of heat; rather, they begin to flow. Subsequent heating/cooling cycles at appropriate temperatures produce thawing/freezing cycles. A thermoplastic molecule is a long-chain molecule and may be thought of as a spaghetti noodle or a hair.
Amorphous thermoplastics (Illustration 1) have random molecular structures, whether we experience them as liquid or solid. Since they have no particular structure, when uncolored they generally are clear and transparent, though many have an amber or brown tint. Since they have no particular molecular structure, they soften over a broad temperature range.
Semi-crystalline thermoplastics (Illustration 2) have zones of crystal structure embedded in an overall amorphous structure. The molecules have an affinity for forming crystals, but because the molecules are so long, only sections of each molecule can join a particular crystal. A fully crystalline material, such as ice, has a very distinct melting temperature, above which the material only can exist in liquid form. Below that temperature, crystal formation occurs and the overall structure "locks" into a particular shape.
There is a significant amount of energy associated with the formation or dissolution of crystal structures, referred to as phase change energy or heat of crystallization that manifests as a more-or-less constant temperature rise as constant power is applied, but at melt temperature the temperature then plateaus until enough heat is absorbed to release molecules from the crystal structures, then constant temperature rise continues. The same occurs in reverse, with a cooling part tending to stay at melt temperature for a while on the way down until it loses the energy it takes to cause crystals to form. Since crystal structures in semi-crystalline thermoplastics scatter light, when uncolored these materials appear milky, translucent or even opaque. If a part made of uncolored semi-crystalline thermoplastic is held against a hot tool, the area that is melting will appear clear – an indication that the crystals have disintegrated – and the resulting molten material exhibits the expected amorphous state.
The ultrasonic welding process
Ultrasonic energy creates heat in thermoplastic parts through intermolecular friction; the incoming compression waves cause the molecules in the plastic to rub together and produce heat. The ability of plastic material to convert mechanical energy to heat is expressed as its loss modulus.
Welding an assembly made of amorphous thermoplastic requires some sort of energy director (Illustration 3), a designed acoustic weak spot that concentrates all of the relative motion between the part vibrated by the horn (sonotrode) and the part held stationary by the fixture. Upon application of ultrasound of sufficient amplitude, the energy director will begin to melt, flowing out from the energy director to provide molten material to create the weld. Contact with cool air is not a particular problem as the time is quite brief and there has been sufficient heat created to promote flow.
When the energy director is fully collapsed and the joint is closed, the ultrasound is turned off and the joint solidifies.
An assembly made of semi-crystalline thermoplastic may be made with an energy director, but because the molten material that begins to flow from the tip of the energy director is immediately removed from the source of heat, it generally solidifies before the joint has a chance to close. For this reason, it is practically impossible to produce a weld wider than the root of the energy director using a semi-crystalline material.
The shear joint (Illustration 4) was created to address that problem. By creating a telescoping zone of molten material, much more material can be involved in the joint and kept from exposure to air, and joints even may be designed that are stronger then the nominal wall of the assembly.
Amplitude is the peak-to-peak distance traveled during one acoustic compression wave cycle. Amplitude must be transmitted from the horn through one of the parts to the joint. Amorphous materials tend to transmit sound waves relatively efficiently and, given the relative ease of joining them, amplitude does not need to be particularly high at the horn. This allows for somewhat complex geometry, relatively large parts and usually more reliable "hermetic sealing" of joints, if required.
Semi-crystalline thermoplastics, by contrast, transmit sound generally much more poorly because the crystal structures tend to scatter and deflect straight-line transmission and the materials often are more "dead" acoustically than amorphous thermoplastics. This often can be demonstrated by tapping parts on a hard surface. Parts made from amorphous materials often will practically ring like a bell by comparison to the often dull thud of a part made of semi-crystalline material. For these reasons, amplitude at the horn usually needs to be quite high, which constrains horn design, which in turn constrains part size and geometry and can cause hermetic sealing to be more difficult.
Polycarbonate, while an amorphous thermoplastic, often is handled as though it were semi-crystalline at the ultrasonic welder. This is because the thermal conductivity and melt temperature of polycarbonate are relatively high. In polycarbonate, energy director welding essentially is a race between destruction of the energy director with clamp force and high amplitude and melting/fusing of the plastic. Polycarbonate parts also can be susceptible to cracking either at the welder or later in the life of the assembly if there is significant molded-in stress. Careful consideration of gate placement and size, vent location, mold temperature, fill speed and pack pressure is especially important for part geometry, weld success and part longevity.
Colorants, additives, fillers, reinforcements and enhancers
Every color added to a thermoplastic material is a chemical that can have its own impact on welding. Color usually is introduced into the part production process as concentrate, which uses a very high melt index (low viscosity when molten) carrier resin with waxes and other lubricants in it to help to evenly distribute the powder or liquid color. Since the principal specification of a color concentrate is its color, its chemistry sometimes can vary enough to cause variations in weld results. At very least, if the parts to be joined come in a variety of colors, adjustments may need to be made at the welding process for each color.
Fillers displace weldable resin with mineral or other non-weldable substances. Sometimes, they are added to increase stiffness and dimensional stability and, to the extent they do so, they usually enhance weldability. Other times, they merely are intended to displace expensive polymer with less expensive material. In either case, in concentrations over 20 percent to 25 percent, these can adversely affect the welding process.
Reinforcements, such as glass fiber or carbon fiber, function like reinforcing bar in concrete, improving tensile strength and dimensional stability in the finished part. Again, to the extent they do so, they improve the welding process. As with fillers, reinforcements displace bondable polymer and, since they generally lie parallel to mold steel, they lie parallel to joint line and tend not to cross the joint plane. So, a reinforced material twice as strong in tensile strength as base resin may have a joint strength potential only 75 percent to 80 percent as strong as base resin or 35 percent to 40 percent as strong as reinforced material.
Fillers and reinforcements generally cause the knit lines (Illustration 5) in parts to be significantly weaker than surrounding material. While knit lines sometimes can cause problems in unfilled or unreinforced parts, problems are much more likely with filled or reinforced parts. The vibrations caused by ultrasonic welding sometimes can find weak knit lines and break them open. In some cases, it is necessary to eliminate the knit lines by removing core pins from the mold cavities and drilling holes as a secondary operation.
Internal lubricants like waxes or stearates ("lubricated" resin) may ease flow in the mold and deliver a clean part release, but lubricants will retard the production of frictional heat in ultrasonic welding. The lubricant also tends to gather at the surface of the part, where it will do the most good in molding but the most harm in welding. It is, in some cases, possible to compensate with welder adjustments.For the same reason, sprayed-on mold release also is undesirable. Any sprayed-on release definitely must be a paintable/platable grade, and silicone-based releases often will prohibit welding completely. Parts thus contaminated often may be made weldable if cleaned with a diluted alcohol mixture. Application of sprayed mold release is, by definition, intermittent and causes a different welding condition on each part and, therefore, a different welding result. It is best to avoid sprayed-on mold release entirely.
Flame retardant in the part also may act like internal lubricant and can make welding more difficult.
Some materials, particularly polyamides (nylon) and – to a lesser extent – polycarbonate, are hygroscopic, meaning they will absorb water from the air. Since water flashes to steam at 100C with considerable phase change energy and polyamide and polycarbonate melt at much higher temperatures, the water trapped in the joint detail will absorb all heat energy created until it can flash to steam, which often leaves a joint detail pounded down by amplitude and clamp force that never got hot enough to melt and weld. If it should occur that the joint does get hot enough to flash the water to steam, then there will be a small explosion occurring within the joint since steam occupies significantly more volume than water (which is the principle upon which a steam locomotive operates). Such joints may appear foamy and brittle and usually lack strength and a "hermetic seal." Usually, a part dry enough to mold is dry enough to weld, so in moderate- to high-humidity environments parts should be welded immediately while still warm or bagged while still warm with a desiccant pack and not unbagged until they are ready to go into the welder.
The question of how well the ultrasonic welding process handles regrind really hinges on two factors: first, how large is the process window for this particular product? Adding regrind to the mix will make the process fussier, which is not a good direction to go if the process already is fussy. If the window is large, some regrind of good quality might not even be noticed. The automotive battery industry makes new battery housings out of 100-percent regrind that comes from the recycling of automotive batteries. The joints are over-designed to produce ultrasonic welding (and other plastic welding) processes with large windows, and everything runs fine. The material also is carefully inspected and property enhancers are added to make certain that the parts and assemblies will be very well made, since a leaky automotive battery is a very bad thing indeed.
The real issue is not how much regrind the process can tolerate, but the quality of the regrind. Regrind from runner systems may be just fine as long as the runner does not see excessive shear heating or damage from excessive pack pressure. Damaged material tends to have a lot of broken molecules in it, expressed as low average molecular weight, which tends to increase the melt index (which make the material runnier) and decreases tensile strength and impact strength. But, perhaps the biggest problem with regrind is thermal damage to the material that turns some of the molecules from bondable polymer into charred filler. Taking great care in the regrind process and monitoring melt flow index of actual parts on a consistent basis likely will head off a lot of problems with the variability that regrind can introduce.
There are many combinations of dissimilar materials that can be ultrasonically welded. In order for a successful weld to occur, the materials must have chemical affinity for one another. Imagine a very low temperature operation involving an attempt to ultrasonically weld a chunk of frozen motor oil to a block of ice – this would be the very definition of an impossible task. Generally, semi-crystalline materials only can be welded to the same material. A few amorphous materials have chemical affinity for more than one other material, but there are not many like this. In addition, the two materials must have similar melt index numbers and similar melt temperatures. The best rule is consult an expert and then experiment thoroughly before committing to a product that requires ultrasonic welding of dissimilar materials.
Knowing the material factors that can affect ultrasonic welding allows for material choices to be made with process impacts anticipated rather than discovered. The story at the beginning of this article unfortunately is one that has been repeated many times and in many places. Discussing these issues can make the ultrasonic welding process seem finicky and sensitive and, if efforts are not made to widen process windows, it indeed can be. However, intelligently applied, the ultrasonic welding process can be robust and stable in both high- and low-volume applications in a wide variety of industries and applications. Making it a successful process simply requires understanding the basic process inputs and making good design and material choices.