“Soldering oil” is a translation problem before it is a technical one. The phrase is used in procurement for at least three unrelated materials — flux, solder mask, and the anti-oxidation oil that floats on a solder pot — and the article that follows treats all three as one. This one separates them, decodes the four-character IPC J-STD-004 flux code that actually appears on a material label, and then adds the part that matters most on a high-frequency board: the residue left behind is 1–10 µm of uncharacterised material sitting exactly where the current flows.
Three Materials Are Sold Under One Name
Before any property is discussed, the term has to be pinned down. In a factory the same two characters appear on three different requisitions:
| Term used | What it actually is | Where it is used | What it decides |
|---|---|---|---|
| Flux / soldering flux (often called soldering oil) | A chemical that removes oxide and lowers surface tension during soldering | Wave, selective, reflow (in paste) and hand soldering | Wetting, joint formation, residue and long-term leakage |
| Solder mask (“solder resist oil”) | A permanent photoimageable polymer layer | Board fabrication, before finish | Insulation, solder dam, and on RF boards a dielectric layer in the field |
| Anti-oxidation oil | A liquid blanket over molten solder | HASL pots, wave solder pots, tinning baths | Dross rate and solder consumption, not joint quality |
The original article used the first meaning throughout and the third in passing, which is why its property list mixes wetting behaviour with pot maintenance. Here the subject is flux — the material that ends up on the board and therefore the only one that can fail in the field. Solder mask has its own article, solder mask on high-frequency PCB.
What Flux Actually Does: Four Jobs, Only One of Them Thermal
Copper in air carries a passive oxide layer, CuO and Cu2O, a few nanometres thick. Molten solder will not bond to it — it balls up and rolls off. Flux has to remove that layer and then keep it from reforming while the joint is molten:
- Oxide removal. Activators — weak organic acids such as adipic, glutaric or succinic acid, or halide salts — become aggressive in a narrow temperature window just below the alloy’s melting point, converting the oxide into a metallic salt that the solder can displace.
- Oxygen barrier. The vehicle — rosin, resin or a synthetic base — melts over the joint and keeps air away during the seconds the metal is bare.
- Wetting. A lower surface tension lets the alloy flow into gaps and around leads by capillary action instead of sitting as a dome.
- Heat transfer. Real, but secondary. In hand soldering most of the heat arrives through the iron tip in contact with the joint and through the solder bridge; flux helps, it does not carry the process.
The temperature window is where flux selection starts to bite. Sn63Pb37 melts at 183 °C and SAC305 at 217–220 °C, and the activator has to be fully consumed by the time the alloy solidifies. Under-activated flux leaves oxide and produces non-wetting; over-activated flux, or a profile that runs cold, leaves unreacted activator in the residue — which is the single most common reason a “no-clean” board later fails a leakage test.
Reading the Four-Character J-STD-004 Code
Flux is not classified by brand or by colour. IPC J-STD-004 assigns a code that states composition, activity and halide content, and it is the only part of a flux label worth transcribing into a build package:
| Position | Values | Meaning | Consequence |
|---|---|---|---|
| 1–2: base | RO / RE / OR / IN | Rosin / resin / organic acid / inorganic | Sets residue character. IN is essentially never used in electronics |
| 3: activity | L / M / H | Low / moderate / high | Cleaning power against oxide, and corrosivity of what is left behind |
| 4: halide | 0 / 1 | 0: no halide detected (copper mirror pass, under 0.5 %) 1: halide present (0.5–2 % at M1, above 2 % at H1) | Halides are the fastest activators and the fastest route to electrochemical migration |
So ROL0 reads as rosin-based, low activity, halide-free — the default no-clean class, and the one most commercial assemblies specify. REL0 is the same activity and halide level on a synthetic resin base, common in solder paste. ORH1 is an organic-acid flux with high activity and halides present: excellent on oxidised or difficult surfaces, and it must be washed off.
One misconception is worth killing here, because it drives wrong cleaning decisions: the code does not tell you whether the board must be washed. OR does not automatically mean water-washable, and RO does not automatically mean no-clean. Whether cleaning is required comes from the cleanliness requirement of the product, the residue’s behaviour under humidity and bias, and whether the assembly gets conformal coating — not from the first two letters. Two fluxes with identical codes from different suppliers can still differ in residue tackiness and cleanability, which is why a substitution should be validated on the line rather than approved on the label.
Residue Sits Exactly Where the Current Flows
On a digital board, flux residue is a cleaning and leakage question. On an RF board it is also an electrical one, and almost nobody treats it that way.
Copper conduction is confined to a skin depth that shrinks with frequency: about 2.1 µm at 1 GHz, 0.65 µm at 10 GHz and 0.39 µm at 28 GHz. A no-clean residue is typically 1–10 µm thick. It therefore sits directly on top of the conductor, inside the region where the field is strongest, and it is a material whose dielectric constant and loss tangent nobody has measured. On a 60 GHz board that is a lossy film a few skin depths thick laid over every trace and pad.
The practical rules that follow are simple. On millimetre-wave boards, wash even when the paste is rated no-clean — the cost of an aqueous or semi-aqueous step is small against a loss budget that cannot absorb an unknown film. Keep flux off the RF launch and connector interfaces, where a few microns changes the transition. And never assume that a residue which passes a bulk ionic test is electrically invisible: ROSE measures conductivity in an extract, not a dielectric layer on a trace.
| Test | Method | What it catches | Commonly specified limit |
|---|---|---|---|
| ROSE / ionic contamination | IPC-TM-650 2.3.25 | Bulk ionic residue extracted from the board | 1.56 µg NaCl equivalent/cm² or tighter by agreement |
| Ion chromatography | 2.3.28 | Individual ions: chloride, bromide, sulfate, weak organic acids | Per-ion limits, used when ROSE passes but coating adhesion fails |
| Surface insulation resistance | 2.6.3.3 | Leakage under humidity and bias | Above 1×108 Ω at day 4 and day 7, 85/85 with bias |
| Electrochemical migration | J-STD-004 annex | Dendrite growth under condensing conditions | Final SIR at or above 10 % of initial after 500 h biased |
Note the gap those tests leave: none of them is a dielectric measurement. A board can pass all four and still carry a film that shifts insertion loss by a few tenths of a dB at 28 GHz. If the loss budget is tight, the only reliable control is process — wash, or keep the residue out of the RF path by design.
Choosing Between No-Clean, Water-Soluble and Rosin
Three chemistries cover almost all production. The decision is driven by what happens after soldering, not by how well they wet:
| Chemistry | Typical code | Residue | Cleaning | Use when | Failure mode if misused |
|---|---|---|---|---|---|
| No-clean rosin / resin | ROL0, REL0 | Benign, tacky, 1–10 µm | Optional; recommended on RF and before coating | Clean copper or OSP, controlled profile, dry environment | Unreacted activator trapped under low-standoff parts; leakage under humidity |
| Water-soluble organic acid | ORH1, ORM1 | Hygroscopic, corrosive | Mandatory, inside a few hours | Oxidised or difficult surfaces, high-yield demand on mixed metallurgy | Residue left in a via or under a connector becomes a field corrosion site |
| Rosin, activated | ROM1, ROH1 | Sticky, removable with solvent | Usually required | Rework benches, through-hole with heavy oxidation | Halide residue with no wash — the classic dendrite recipe |
The cost comparison usually runs the other way from intuition: the cheap flux is the one that forces a wash step, a DI-water skid, a bake and a cleanliness record. No-clean exists because it removes a process, not because it solders better.
PTFE and Low-Energy Surfaces: Where Flux Stops Working
On a PTFE board the wetting problem is different in kind. PTFE surface energy is around 18–20 mN/m, far below the 40 mN/m order of a solder-mask or FR-4 surface, so molten solder has almost nothing to grip. A bare PTFE laminate cannot be soldered at all: a metallic finish is mandatory, and the finish is what the flux actually acts on.
- The finish decides the flux. OSP and immersion silver take a mild ROL0 or REL0 cleanly. ENIG needs less activity, not more — and on RF boards the nickel under the gold is already inside the skin depth at millimetre-wave frequencies, so the finish choice belongs to the loss budget before the flux does (see surface treatment on high-frequency PCB).
- Residue does not spread or rinse the same way. On a low-energy surface the flux tends to bead and stay where it landed, including in the gap under a component, which makes a late wash much less effective. Wash soon after reflow, before the residue hardens.
- Hybrid assemblies ask the flux to serve two surfaces. Most 5880 and TLY-5 designs are hybrids with an FR-4 or 4350B section for control and power, so the same flux meets a solder-mask surface and a PTFE edge in one pass. The PTFE parameters are in the 5880 laminate data and the TLY-5 process notes.
- Conformal coating is not a substitute for washing. Coating bonds to residue rather than to laminate or mask, and lifts later in sheets. Wash and bake to J-STD-033 before coating, every time.
Six Flux Failures That Reach the Customer
All six pass final test. That is what makes them worth listing:
| Failure | Mechanism | Where it surfaces |
|---|---|---|
| Electrochemical migration | Ionic residue plus moisture plus bias grows a dendrite between adjacent conductors | Months later, as an intermittent short after a humid season |
| Leakage on high-impedance nodes | Residue film conducts weakly when wet; a 1012 Ω node becomes 109 | Sensor and front-end drift, calibration failures |
| Coating delamination | Adhesion to residue rather than to the board | Coating lifting in sheets at the edges and around tall parts |
| Trapped flux under low-standoff parts | QFN, LGA and connectors shield the residue from any wash that arrives late | Corrosion under the part, invisible until cross-section |
| Charred residue after double reflow | Lead-free profile plus a second pass cooks a no-clean residue into a hard brown film | Cosmetic rejection, and poor probe contact on test pads |
| Loss and detuning on RF | An uncharacterised dielectric film over the conductor | Insertion loss a few tenths of a dB high, matching never quite closing |
What Belongs in the Assembly Note
Five lines turn flux from a consumable into a controlled material:
- The J-STD-004 designator, not the brand: “no-clean, ROL0 or REL0, halide-free.” A brand name lets a substitution through; a designator does not.
- The cleaning decision, stated as a requirement: “no-clean permitted” or “wash required, ROSE ≤ 1.56 µg/cm² on 100 % of units.” Silence is read as permission to skip.
- The cleanliness test and sample rate, named. A requirement without a test is a wish.
- The coating sequence: wash, bake to J-STD-033, then coat, with the adhesion test specified.
- RF-specific: on boards above roughly 20 GHz, state that no flux residue is permitted in the RF path and on connector launches, and require a wash even with no-clean chemistry.
The same discipline applies at the fabricator’s end, where flux appears again in HASL and in the final finish line; the finish and its thickness tolerances are covered in the surface finish comparison.
Five Claims in the Original Article That Do Not Hold
| Claim | What is actually true |
|---|---|
| “Soldering oil is also commonly referred to as flux” | Partly. The term covers flux, solder mask (“solder resist oil”) and the anti-oxidation oil on a solder pot — three materials with nothing in common. Only flux ends up on the finished board |
| “It accelerates heat transfer from the iron tip” | A minor effect. Flux removes oxide, blocks re-oxidation and lowers surface tension; heat transfer in hand soldering comes mostly from tip contact and the solder bridge |
| “It protects the base material and returns to protecting the joint after soldering” | No-clean residue is not a protective layer. It is hygroscopic and ionic, and under bias and humidity it is the path that grows dendrites. Protection comes from the finish and the coating |
| “Modern formulations are environmentally friendly” | Halide-free is not the same as low impact. Water-soluble fluxes load the waste stream with organic acids and demand treatment; the honest comparison is per litre of make-up and per litre of effluent |
| “Flux reduces cold joints and false soldering” | Flux prevents non-wetting. Cold joints come from the thermal profile, cooling rate and joint geometry; a cold profile with the best flux still produces a grainy, weak joint |
Conclusion
Three sentences carry the subject. The word covers three different materials, and only flux is a board-level material. The four-character J-STD-004 code is the only part of a flux label worth specifying, and it does not tell you whether to wash. And on an RF board the residue is not chemistry but a dielectric layer of unknown properties sitting on the conductor — which is why the correct answer at millimetre wave is to wash regardless of what the paste is rated.
Practically: name the designator, state the cleanliness test and its limit, wash before coating, and treat the finish and the flux as one decision rather than two.
XCE PCB builds high-frequency assemblies on FR-4, RO4003C, RO4350B and PTFE laminates from 2 to 72 layers, holds impedance to ±10 %, and documents the finish, the cleanliness test and the coating sequence per order — samples in as fast as 24 hours. The twelve-step production flow that these steps belong to is in the manufacturing process overview.
Frequently Asked Questions
Is soldering oil the same as flux?
What does ROL0 mean on a flux label?
Does a no-clean flux mean the board never needs washing?
How clean is clean enough?
Why is flux residue a problem on high-frequency boards specifically?
Can you solder directly to PTFE laminate?
Need a Board Finish and Flux Pairing That Survives RF Testing?
Send us your Gerber files and requirements for a manufacturability review and a tailored quotation.
sales1@xcepcb.com
+86 180 3341 4657