Published: January 20, 2026
This blog post discusses the process of removing semiconductor components from a printed circuit board (PCB).
One day, Senpai* C, a veteran employee, came to Rookie D, in her second year with the company…
* Senpai: A Japanese-style title to show respect to those older than oneself, especially used in organizations such as companies, schools, etc. It sometimes implies that the psychological relation between the two is closer than that between a boss and a subordinate.
Are Semiconductor Components Returned in Damaged Condition?
To Minimize the Impact of Moisture Inside Semiconductor Components
Water Expands by 1700 Times in Volume When It Boils and Turns to Steam!
Package Crack Can Lead to Corrosion
Bonding Wire Damage Caused by Package Crack
If Possible, Remove Them by Reflow!
Help Me, Senpai! Series Vol. 9
How to Safely Remove Semiconductor Components—Handle with Care
Today, I’d like to talk about something a bit different, which is how to remove semiconductor components.
What’s going on?
I’ve been told by the analysis team that they’re having difficulty with the failure analysis of semiconductor components.
What kind of difficulty?
It seems semiconductor components are returned in damaged condition.
Well, if they’re defective, isn’t it normal they’re damaged?
No, I don’t mean that. They seem to get damaged during removal from PCB.
Surely they’re not being forcibly removed with nippers or pliers, right?
That’s true, but it seems they’re being removed using a soldering iron.
I just assumed that was how it’s usually done.
I see… So that’s considered normal, then…
Is that not correct?
You know how to mount semiconductor components, don’t you?
Yes, they’re usually mounted using reflow soldering, right?
That’s right. Let me explain a bit more. If the semiconductor components have absorbed moisture, they’re first baked to remove the moisture. Then comes the reflow process, where the temperature gradually increases, and the peak temperature is maintained only briefly. This is specified in what’s called a “reflow profile.” Do you know what that means?
I just assumed that’s how it’s usually done.
This process is intended to minimize the impact of moisture inside semiconductor components.
Semiconductors are encapsulated in epoxy molding compound, but moisture can still penetrate through the compound itself or at the interface between the leads and the molding compound.
Ah. I think it’s called MSL, right? Moisture sensitivity level, is it?
Exactly. Moisture sensitivity level. It’s an index that indicates how susceptible a semiconductor component is to moisture absorption. The higher the number, the greater the sensitivity. Generally, larger packages tend to have higher MSL ratings. Here’s a simplified version of the MSL table.
| MSL | Floor life*1 |
| 1 | Unlimited*2 |
| 2 | 1 year |
| 2a | 4 weeks |
| 3 | 168 hours (1 week) |
| 4 | 72 hours (3 days) |
| 5 | 48 hours (2 days) |
| 5a | 24 hours (1 day) |
| 6 | Baking is required before mounting. |
*1 Floor life: Storage time in room condition. The maximum allowable time that components can be exposed to ambient room conditions after being removed from moisture barrier bags, prior to mounting. The conditions are ≤30°C and ≤60% relative humidity.
*2 Unlimited condition: ≤30°C and ≤85% relative humidity
Yes. What about it?
Why does moisture sensitivity matter? You see, water boils at 100°C, right? So, what do you think the typical reflow temperature is?
Something like 250°C, I think.
Right. So, what happens if moisture remains inside the component and boils at that temperature?
It would turn into steam and escape, right?
But since it’s encapsulated in epoxy molding compound, there’s no way for the steam to escape. When water turns into steam, its volume increases by roughly 1,700 times. If that much steam has nowhere to go, what do you think happens? The epoxy molding compound can crack—or in the worst case, the component can burst like popcorn.
Photo 1. Example of package crack
Ah, so that’s how package crack happens. That’s why we have a reflow profile—to prevent that. By raising the temperature slowly, moisture inside the package gradually evaporates, similar to baking.
Exactly. But you still need to respect the floor life. The reflow profile isn’t a substitute for baking, since baking is done over several hours.
By the way, do you know the temperature of soldering iron?
Well, it’s definitely above 100°C, I think.
No, it can even be as high as 350°C. Since the reflow profile peaks around 250°C, that’s significantly higher, isn’t it?
Yes. That’s a bit too high.
Apparently, the temperature is set a bit higher to account for cooling, but what do you think happens when you use this soldering iron to remove a semiconductor component from a PCB?
Since it absorbs moisture even while mounted on a PCB, would sudden application of a hot soldering iron cause the package to crack?
Yes. That can happen, and sometimes small gaps form between the leads and the molding compound, allowing steam to escape through those gaps.
I see… so that means…
A large crack in the package would be obvious, but small gaps between the leads often go unnoticed, allowing moisture or other chemicals to enter and corrode the IC chip.
| Non-corroded IC pad | Corroded IC pad |
Photo 2. Example of pad corrosion (the circular features are bonding wires)
Does corrosion mean it rusts?
Yes. Of course, ordinary moisture can cause corrosion, but chemical contaminants are even more problematic. Salt is a common example. Near the sea, salty air can get in and cause corrosion. In hot spring areas, various gases are released, and if it’s cold, condensation can let moisture in. There are many factors.
Typically, it takes no more than a week to ten days to remove components from a PCB and send them to the semiconductor manufacturer. Is it really possible for corrosion to occur in such a short period?
Exactly. The problem is that it’s impossible to tell whether the corrosion was caused by an original defect or after the component was removed.
Can’t it be determined, like figuring out the time of death in a crime drama, whether it happened before or after?
…….
Sorry…I see, so it can’t be determined.
In that case, the quality assurance team ends up reporting something like: “Failure due to corrosion. It is suspected that corrosion occurred due to moisture entering through a small crack caused by heat during the assembly of the printed circuit board.” There might actually be another cause for the defect, but if that can’t be clarified, it’s not really resolved for either the semiconductor manufacturer or the customer. The risk of recurrence remains high.
I see. So, it’s actually being damaged during removal from a PCB.
Like I said, if the true cause can’t be clarified, it’s not really a resolution for either the semiconductor manufacturer or the customer. The risk of recurrence remains high.
Moreover, the presence of cracks in the package indicates that the molding compound is under considerable stress. Stress here refers to forces that deform the molding compound. That can also damage the bonding wires connecting the IC chip to the lead frame.
Photo 3. Example of a detached bonding wire (corrosion also present)
So that would result in an analysis report stating, “bonding wire open,” right?
No, no, that would be very unlikely. That’s a different failure mode. At that point, it becomes clear that something is off. With the wire detached, the component wouldn’t function at all, so it clearly doesn’t correspond to the abnormal operation that was reported.
I see. But how exactly should semiconductor components be removed from a PCB? Even when I search online, the results say, “Use a soldering iron to remove it.”
Well, online instructions don’t assume the removed components will be evaluated or analyzed, so they don’t worry about potential damage. To remove them without damage, reflow is the right approach.
I see, so it can be removed using the same method as when it was mounted. Just heat it gradually and remove it before it cools. But are reflow systems readily available?
Since components can be mounted by reflow, such equipment exists. But you can’t just take the PCB to the production line for removal. In a lab environment, I’d expect there to be a compact reflow system… If that’s not available, heat the IC gradually, following a profile similar to the reflow profile, to properly remove moisture inside the package.
I’d like to know the exact method.
Specifically, for removing bottom-terminated ICs like CSPs or BGAs, use a localized heating tool. It looks like a hair dryer and blows hot air—sometimes called a heat gun. Of course, it’s better to use a heat gun that allows temperature control, so that high-temperature air isn’t applied directly to the IC package. If you keep the package surface temperature below 200°C, you can minimize damage from absorbed moisture. You know, even at 200°C, the solder melts, so you can still remove the IC.
I see. That makes sense. Thank you.
Let me explain in a bit more detail.
Yes?
The issue with using a soldering iron is that it doesn’t gradually heat the entire IC package. Instead, it rapidly heats only the leads to around 350°C. This quickly transfers heat from the metal leads to the lead frame, bonding wires, and the IC chip. But the epoxy molding compound isn’t heated yet. That creates a large temperature gradient.
I see. And then?
As a result, the plastic molding compound is rapidly overheated in localized areas, causing moisture to vaporize and expand rapidly in volume. This can generate stress, which may result in cracks in the package or delamination between the bonding wires and bonding pads.
Figure 2. Mechanism of package crack and delamination between bonding wire and bonding pad
On top of that, mechanical stress from the soldering iron combines with thermal stress, causing damage to the semiconductor components.
I see. That makes things a bit clearer.
By the way, solder used in the semiconductor and PCB assembly industry is mainly Sn3Ag0.5Cu, with a melting point of about 217 to 220°C. You might think 200°C would not be sufficient to melt it. However, even if the temperature of the epoxy molding compound is around 200°C, the leads themselves can exceed 200°C. This is because they are metallic and have high thermal conductivity. As a result, the solder melts without any issue.
Understood. Thank you very much.
Learn more about quality and reliability at Nisshinbo Micro Devices here.
Learn more about reliability test specifications (electronic device products) at Nisshinbo Micro Devices here.
Afterword
This is the ninth installment in our blog series introducing how our products can help you.
If you have any questions about this blog, feel free to contact us below.
Click here to read the previous blogs in the “Help Me, Senpai!” series:
Vol. 1: IoT for Energy Harvesting: Solar Panel Can’t Drive System till Morning
Vol. 2: Stop Sleeping Mobile Device from Consuming Battery Charge!
Vol. 3: RTC Backup Switchover Circuit Is Not Easy.
Vol. 4: What Is AEC-Q100?
Vol. 5: Buy Samples Online - No Sales Pressure Involved
Vol. 6: How do I Use This Type of Reset IC? (What is Functional Safety?)
Vol. 7: Why Not Just Eliminate Flicker?—Exploring Analog (Linear) Dimming
Vol. 8: Quality Assurance for Semiconductors—and What’s the Bathtub Curve?
Blogs about quality assurance and reliability:
Vol. 4: What Is AEC-Q100?
Vol. 8: Quality Assurance for Semiconductors—and What’s the Bathtub Curve?