The Sample Is Not Neutral: How Extraction Changes The Result

Overview

A cleanliness measurement is usually reported as a single figure in milligrams, and it is discussed as if that figure were a property of the part. It is not. It is a property of the part and the method, and the method contributes more variation than most people expect.

Three laboratories given the same component and asked for its residual mass will produce three different numbers if they use different extraction methods, and each will be correct within its own procedure. The disagreement is not error; it is three different questions being answered and reported as if they were the same one.

This article sets out what each common method actually extracts, where the variation comes from, and how to write a sampling procedure that produces a number two parties can compare. Big Bird Industrial (formerly Shimada Big Bird Industrial) has been building cleaning and deburring systems for automotive powertrain components since 2001, and the methods described here are the ones that appear in production acceptance specifications.

1. Three Ways To Sample

The extraction method determines how much of the contamination is available to be measured, and the three common methods differ substantially in what they release.

Wiping

A defined area is wiped with a specified cloth or swab, and the material collected is measured. Wiping is fast and portable, and it recovers what is loose on a surface. It recovers very little of what is inside a cavity, behind a feature or in a thread root, which means a wiped result describes the visible surface rather than the part.

Flushing Or Rinsing

Liquid is passed over or through the part and the effluent is filtered and weighed. Flushing reaches features that a wipe cannot, and it is the method most often used for passages. Its efficiency depends on flow rate, direction and duration, which is why the procedure has to specify all three rather than simply saying the part is rinsed.

Ultrasonic Extraction

The part or a section of it is immersed and cavitated, which releases material from pores and blind features that flow alone does not reach. It is the most aggressive of the three and generally returns the highest figure, which makes it useful for establishing what is actually on a part, and less useful for routine acceptance where speed matters.

2. Why The Numbers Diverge

The differences between methods are not small, and understanding the mechanism explains why comparing them is meaningless.

Access

A wipe contacts the surfaces an inspector can reach. A flush contacts the surfaces liquid can flow across. Cavitation contacts the surfaces liquid can wet, including inside pores. Each step increases the accessible area, so each step tends to increase the recovered mass on the same part.

Attachment Strength

Wiping removes what is resting on a surface. Flushing removes that plus what is loosely adhered. Cavitation removes that plus what is held in fine features by surface tension and geometry. A part can measure clean by wipe and fail by ultrasonic, and both results are honest descriptions of different questions.

3. The Solvent Question

The liquid used for extraction is part of the method, not a neutral carrier.

Water Or Solvent

Water removes water-soluble material well and leaves oil-based contamination largely in place. A solvent removes oils and leaves some inorganic salts. A method that uses water will report a lower figure on an oily part than one that uses solvent, and neither is wrong; they are measuring different fractions.

The Drying Step

After extraction the filter is dried and weighed, and the drying condition affects the result. An incompletely dried filter weighs heavy; an overheated one can lose volatile material. Stating the drying temperature and time is as important as stating the solvent.

4. Where You Sample

On a part with several regions, the location of the sample determines the result more than the method does.

The Region Defines The Question

A figure taken from a machined face describes a machined face, and a figure taken from inside a gallery describes a gallery. Reporting either as the cleanliness of the part implies a uniformity that does not exist, and on a head the spread between regions can be two orders of magnitude.

The Sampling Plan Is Part Of The Specification

Which parts, which regions, how many samples and at what interval are all decisions that belong in the acceptance document. A specification that names a mass limit without naming the region and the frequency leaves the measurement to whoever performs it.

5. Who Does It, And Does It Matter

It matters, and the variation is measurable.

Operator Effect

Wiping especially depends on pressure, speed and pattern, all of which vary between people. Flushing depends less on the operator, which is why it is preferred where the result has to be defensible.

Between Laboratories

Two accredited laboratories following the same written procedure can still differ, because a procedure leaves room for interpretation at the boundaries: how long the part drains, whether the filter is pre-weighed, how the sample is transported. The practical answer is to exchange samples between the two before a dispute rather than after.

6. Making It Repeatable

A sampling procedure that produces comparable numbers specifies six things: the region, the method, the liquid, the flow or cavitation parameters, the drying condition, and the balance resolution.

The balance resolution is worth naming explicitly. A limit of two milligrams measured on a balance that reads to one milligram carries an uncertainty of half the requirement, which makes the result unusable for a decision regardless of how carefully the extraction was performed.

7. Comparing Two Suppliers

The most common failure of comparison is two quotations or two reports using different methods and treating the numbers as equivalent.

Before comparing, confirm that the region, the method and the liquid match. Where they do not, the only valid comparison is a round-robin: the same parts, measured by both parties or by a third laboratory using one agreed method. That exercise costs a week and settles what a year of correspondence will not.

8. A Worked Comparison

Three methods applied to the same machined housing, in the same laboratory, produce a spread that is larger than most acceptance windows.

The Numbers

A wipe over a defined face might recover 0.5 mg. A flush of the same face and its adjacent passage might recover 2 mg. Ultrasonic extraction of a section might recover 6 mg. All three describe the same part, and a specification written at 2 mg would pass it, fail it or pass it depending on which method the laboratory happened to use.

That spread is not a laboratory error. It is the difference between sampling a surface, a surface plus a passage, and a volume including pores. The figure is only meaningful with the method attached, which is why a specification that quotes milligrams alone is quoting half a requirement.

The practical consequence for a supplier is that quoting against a limit without a method is quoting against an unknown. The consequence for a buyer is that a supplier reporting a low figure may simply be measuring less.

9. What A Supplier Can Defend

A supplier asked to meet a cleanliness figure has one strong defensive position and one weak one, and knowing which is which prevents a contract that cannot be met.

The Strong Position

A stated method, a stated region and a stated sample size, priced accordingly. That offer can be verified, repeated and defended in a dispute, because it describes exactly what will be measured and how.

The Weak Position

A figure without a method. It invites the buyer to measure however it prefers, which in practice means the strictest available interpretation, and it makes every acceptance a negotiation rather than a check. A machine capable of 2 mg by flush is not the same machine as one capable of 2 mg by ultrasonic, and the price difference between them is substantial.

Refusing to quote without a method is not obstruction. It is the only way to quote a number that can be delivered.

Frequently Asked Questions

Which Method Is The Right One?

The one that matches the risk. A fluid circuit needs flushing; a visible surface can be wiped; a reference measurement or a failure investigation justifies ultrasonic.

Why Does My Supplier Report A Lower Figure?

Almost always because the method, the region or the liquid differs. Ask for the three, and the discrepancy usually explains itself.

Does Ultrasonic Overstate Contamination?

It reports more, because it reaches more. Whether that is overstating depends on whether the parts it releases would have come loose in service.

How Many Samples Are Enough?

Enough to show the spread rather than a single point. Three samples from the same batch tell more than one sample measured three times.

Conclusion

A cleanliness figure is a property of the part and the method together. Wiping, flushing and ultrasonic extraction answer different questions, and reporting any of them as the cleanliness of the part overstates what was measured.

The remedy is a procedure that names the region, the method, the liquid, the parameters, the drying condition and the balance resolution. Where two parties cannot agree on a method, a round-robin on the same parts resolves in a week what correspondence cannot resolve at all.