Written and maintained by CASRAI Editorial Board
Published · Last updated
On this page
Flip chip is a semiconductor interconnection method in which a die faces downward and connects to a carrier through conductive bumps on its active surface. It replaces the wire loops between die and package. Understanding it requires separating the die connections, the finished package and the evidence used to assess their reliability.
Ask CASRAI · free to tryQuestion about this guide? Ask CASRAI answers with linked sources. 2 free questions a day, no account.
Short answer
The name describes the orientation and connection of the die. In the IEEE Technology Navigator explanation, the active face points toward the substrate rather than upward as in conventional wire bonding. Connections cross the gap through bumps. The historically associated abbreviation C4 means controlled collapse chip connection.
Flip chip therefore identifies an interconnect approach, rather than a transistor architecture or a complete specification for a finished component. A useful distinction when reading packaging literature is to ask which layer a term describes: the die connection, its carrier, or the external connections to a circuit board. These descriptions can all apply to the same assembly without being interchangeable.
Definition and scope
For example, Texas Instruments’ cross-section of a flip-chip ball grid array shows an internal flip-chip bump and a separate external solder ball. The first connects the die to its package substrate; the second is part of the package’s board connection. This distinction explains why a report about BGA board assembly does not automatically describe die attachment.
IPC-7094A’s scope includes bare-die and die-size components, including wafer-level BGA, in direct chip attach assemblies. The document connects component characteristics to assembly methodology. That scope identifies a practical boundary: describing the connection concept is different from specifying the assembly requirements for a particular component format. A label needs its construction context before it can support that more specific discussion.
Integrated circuit design provides the upstream context for the chip itself. This guide focuses on the packaging connection and its interpretation in assembly and reliability work.
How flip chip works in practice
Bumps replace wire loops
ASE’s packaging overview describes solder or gold bumps connecting the flipped chip to a substrate or leadframe. Its wafer-bumping description places bump formation before the wafer is divided into individual chips. With this approach, connection pads can occupy an area across the die instead of being restricted to its perimeter.
ASE links the shorter circuit path and absence of bonding wires to reduced signal inductance. It also describes the opportunity to place a heat sink directly against the chip’s exposed rear surface. These are architectural advantages; they do not specify the performance of every component using the architecture. The distinction matters when a general packaging description appears beside a particular device’s electrical or thermal data.
The interface includes more than solder
IEEE describes under-bump metallurgy as providing adhesion, a diffusion barrier and a surface that solder can wet. In its solder-based example, aligned connections are joined during reflow, and the solder geometry changes as it melts. The supporting metallization is consequently part of the interconnect, even though a simple diagram may show only a bump.
Amkor’s copper-pillar description illustrates the range within that interface. It lists standard pillars, fine-pitch pillars, copper-bar structures and microbumps, with different metal stack-ups for different applications. Copper pillar belongs within the flip-chip vocabulary; it does not by itself identify the entire package.
Underfill supports the connection mechanically
Nordson defines underfill as epoxy occupying a gap between a chip and carrier, or between a finished package and its PCB. It reduces strain associated with unequal thermal expansion and helps protect solder connections against mechanical loading. It is distinct from the conductive interconnect.
Nordson distinguishes capillary material that flows underneath a component, no-flow material present before attachment and reflow, and molded underfill. The terms describe different approaches to filling and reinforcing the gap. They should not be read as different names for one universal material or assembly sequence.
Types and related concepts
FCBGA and FCCSP describe package formats
ASE ChungLi describes FCBGA as combining flip-chip technology with a BGA substrate. Its FCCSP description also combines flip-chip interconnection with a BGA substrate, emphasizing a compact package and molding that encapsulates the die. The abbreviations add package context to the underlying interconnect method.
Amkor’s FCBGA offering shows why the full construction matters. It includes laminate or ceramic substrates, multiple routing layers and vias, and bare-die or lidded package formats. Its bump options include solder and copper pillar. An FCBGA description alone therefore leaves several important construction choices unspecified.
Amkor also lists BGA footprint pitch separately from die bump pitch. When comparing two descriptions, keep the interface attached to the measurement. A fine die connection and a coarser external board connection can coexist; comparing their pitches as if they describe the same joint produces a misleading comparison.
Chip-on-wafer describes a different integration level
ASE ChungLi describes chip-on-wafer as bonding individual dies to a silicon interposer or another logic wafer through microbumps before subsequent packaging. This adds information about where integration takes place. It is useful context when reading about heterogeneous assemblies, rather than a reason to treat every flip-chip component as a stacked package.
The data table below brings these terms together across the manufacturers’ descriptions and IEEE’s terminology. It separates the kind of connection from the package or integration level. Its purpose is to interpret technical language, not rank the technologies or establish interchangeability between parts.
Where flip chip matters for packaging labs
A detector-research example
In Pixel detector hybridisation with Anisotropic Conductive Films, J. V. Schmidt and colleagues report a single-die interconnection approach developed within CERN’s EP research and development programme and the AIDAinnova collaboration. Their application concerns sensor and readout-chip integration for detector research, where small-scale work needs to accommodate individual dies from multi-project wafers.
The authors describe conductive particles in adhesive film replacing solder bumps, with the connection formed using a flip-chip device bonder. Their reported Timepix3 assemblies use a 55 μm pixel pitch and 14 μm pad diameter. These are dimensions of that research example, not general flip-chip design limits. The example shows why the name of a bonding tool alone does not specify the interconnect material: the actual connection structure and application still matter.
Electrical, thermal and materials questions intersect
ASE ChungLi lists high-performance computing, RF applications and automotive electronics among flip-chip uses. It explains that eliminating wire-loop space supports compact packages, while the upward-facing die backside offers access for cooling. These characteristics make the package connection relevant alongside the circuit’s intended function.
IEEE identifies a countervailing mechanical issue: the die and an organic substrate expand differently with temperature, placing cyclic stress on their joints. Short electrical connections do not remove this materials problem. Packaging performance has to be understood across electrical and mechanical behavior rather than inferred from connection length alone.
Henkel’s June 2024 underfill whitepaper ties material effectiveness to interconnect pitch, gap height and processing preference. It explains that flow and gelation behavior must allow complete coverage under large dies. Cured protection and the ability to fill the intended geometry are separate material considerations.
Henkel also discusses warpage and die cracking associated with thermal and mechanical stress, and explicitly recommends application trials because materials and working conditions vary. Its discussion supports evaluating a material in the intended construction. It does not establish that a named underfill is suitable for every flip-chip assembly.
A checklist for interpreting test evidence
Texas Instruments’ SPRU811A reference, dated May 2005, provides a useful distinction between electrical paths and inspection targets. The following checklist translates that distinction into questions for reading a laboratory report. It does not supply a test procedure or acceptance limit.
- Identify the connected path. TI’s daisy-chain example includes die metallization, bumps, package traces, solder balls and PCB traces. Its continuity result concerns the connected circuit.
- Identify the visible feature. TI discusses X-ray inspection for solder-joint geometry and faults, while noting that some defects remain difficult to discern.
- Identify the material interface. TI describes scanning acoustic microscopy as imaging internal features, including voids and delamination.
- Separate location from cause. TI describes time domain reflectometry as helping locate a failure region. A location is a different result from an image of damaged material.
The practical reading is to match each conclusion to the evidence that supports it. An electrical result, an image and a materials observation answer different questions. A report becomes easier to assess when it names the assembly level and the feature examined instead of using a single broad label such as package quality.
Surface mount technology gives the adjacent board-assembly context. For researchers working with optoelectronic assemblies, photonic integrated circuits provide another application context; the connection architecture still needs its own description.
Standards context: scope and maintenance status
IPC-7094A, January 2018, is titled Design and Assembly Process Implementation for Flip Chip and Die Size Components. Its published scope addresses direct chip attach, including design, assembly methodology, inspection, repair and reliability. It identifies managers, engineers, operators and technicians involved in electronic assembly as its audience.
The official IPC standards revision table lists IPC-7094 as No Longer Maintained. The table defines that status as not expected to enter revision at present, while allowing future revision or reaffirmation. The status should therefore accompany a reference to this document; it should not be presented as an actively maintained rulebook merely because its subject fits the component.
This separates two useful questions when reading a technical reference: what assembly does the document address, and what is its publication status? The title answers neither question completely. A laboratory report can identify its exact reference without turning that reference into a claim about every packaging technology discussed here.
Common misunderstandings
A package label specifies the whole construction
It does not. Amkor’s FCBGA description includes different substrate materials, bump choices and lid configurations. For interpretation, retain those qualifiers beside the package label. Two records that both say FCBGA need not describe the same construction, even when their broad packaging category agrees.
A bumping capability guarantees a conversion
Amkor describes copper-pillar compatibility with bond-pad arrangements designed for wire bonding, along with wafer-level electrical testing before pillar bumping. That is a statement about its technology offering. It is not a general guarantee that any existing wire-bond die can be converted without considering its particular design.
Any underfill reference concerns the die
Nordson explicitly distinguishes chip-to-carrier underfill from board-level underfill between a completed package and PCB. A material description needs that location to be meaningful. Otherwise, a statement about reinforcement under a finished component can be mistaken for a statement about its internal die connections.
One inspection proves the complete assembly
TI’s reference treats electrical testing and imaging as complementary forms of evidence. Its examples do not turn a continuity reading or one inspection image into proof of every internal interface. The appropriate conclusion stays tied to the path or feature examined and the documented scope of the result.
Sources
- IEEE Technology Navigator: Flip chip. Checked 2026-11-14.
- ASE: Flip Chip Packaging. Checked 2026-11-14.
- ASE ChungLi: Flip Chip Packaging. Checked 2026-11-14.
- Texas Instruments: Flip Chip Ball Grid Array Package Reference Guide, SPRU811A, May 2005. Checked 2026-11-14.
- Amkor: Copper Pillar. Checked 2026-11-14.
- Amkor: FCBGA. Checked 2026-11-14.
- Nordson: Underfill. Checked 2026-11-14.
- Henkel: Ensuring Large Die Durability in AI and HPC Advanced Packages, June 2024. Checked 2026-11-14.
- Schmidt et al.: Pixel detector hybridisation with Anisotropic Conductive Films, arXiv:2210.13046. Checked 2026-11-14.
- IPC-7094A: Published scope and contents, January 2018. Checked 2026-11-14.
- Global Electronics Association: IPC Standards Revision Table. Checked 2026-11-14.
Data: Flip-chip terminology by assembly level
Original qualitative crosswalk compiled from IEEE, Amkor and ASE ChungLi descriptions. It separates an interconnect method or structure from package formats and chip-on-wafer integration. Entries describe the cited technologies, not interchangeable component specifications.
| Term | Description level | What the term identifies | Source |
|---|---|---|---|
| Flip chip / C4 | Die interconnection | Face-down die joined through conductive bumps; C4 means controlled collapse chip connection. | IEEE Technology Navigator |
| Copper pillar | Interconnect structure | Copper-based bump technology used in flip-chip interconnects, with multiple structures and metal stack-ups. | Amkor: Copper Pillar |
| FCBGA | Package format | Flip-chip BGA package with substrate routing; Amkor lists laminate or ceramic substrates and bare-die or lidded formats. | Amkor: FCBGA |
| FCCSP | Package format | Flip-chip chip-scale package; ASE ChungLi describes a BGA substrate and molding around the die. | ASE ChungLi |
| Chip-on-wafer | Integration architecture | Dies bonded through microbumps to a silicon interposer or logic wafer before subsequent packaging. | ASE ChungLi |
Sources
- IEEE Technology Navigator: Flip chip (accessed 2026-11-14)
- Amkor: Copper Pillar (accessed 2026-11-14)
- Amkor: FCBGA (accessed 2026-11-14)
- ASE ChungLi: Flip Chip Packaging (accessed 2026-11-14)
Compiled by CASRAI from the sources above. Free to reuse under CC BY 4.0 with attribution.







