Start with the question the test must answer
A supplier sends a soft fabric labelled “bamboo blend”. The buyer wants to confirm its fibre content before approving the order. A burn test suggests a cellulosic material, while an infrared result shows cellulose. Has the laboratory proved the label?
Not yet. The results describe part of the material's identity. They do not establish the plant source, the blend percentage or the supplier's brand claim. Each question needs evidence that can answer it.
Textile fibre identification works best as a series of linked checks. First, the laboratory defines the sample and the question. It then chooses tests that can tell the likely materials apart. The aim is a sound answer, with a clear limit to what the evidence proves.
A quick test can be useful. Errors arise when its result is used to answer a broader question than the test can resolve.
Separate identity, quantity and origin
Three claims that sound similar need different kinds of proof:
Question | Example conclusion | Evidence needed |
|---|---|---|
Which fibre types are present? | Cotton and polyester were found | Suitable qualitative tests on fibres that represent the sample |
How much of each is present? | The dry fibre blend contains a stated mass share of each | A validated quantitative method with a defined basis |
Where did the material come from? | The fibre came from a named producer or feedstock | Evidence specific to that origin claim |
AATCC TM20-2021 covers qualitative identification through physical, chemical and microscopic techniques. It separates that task from measuring blend percentages. AATCC TM20A-2025 covers quantitative analysis, including moisture, nonfibrous matter and fibre composition.
A laboratory may need both types of work. Finding two fibre types does not reveal their mass ratio. To measure the mass ratio, the analyst needs a suitable method. Knowing which fibres are present helps with that choice.
Origin is a further question. Finding polyester does not name its maker. A cellulose result does not, by itself, establish the tree or plant from which the pulp came.
Map the textile before taking it apart
Start by examining how the sample is built. Warp and weft may contain different materials. So may the face and back, the pile and ground, or an elastic core and its covering. A small sample from one part may not represent the whole.
Consider an imagined coated stretch fabric. Label the face yarn, back yarn, elastic core and coating as separate parts. Keep their locations clear as they are removed. Mixing them may still allow several polymers to be found. It can hide where each one occurs.
The sampling plan should match the claim. To identify a sewing thread, test that thread. For the main fabric, check what the agreed method includes. An attached thread or label should not enter its fibre blend unless the method requires this.
Keep an untouched piece where enough material is available. It allows a fresh test if an odd result appears. This spare piece must come from the same sample. A similar-looking swatch from another roll cannot replace it.
A burn test narrows the possibilities
Burn behaviour may help distinguish broad groups of fibres. It is much less useful as a stand-alone answer for a blend or a heavily treated textile.
The Canadian Conservation Institute's identification guidance treats burning and microscopy as supporting methods. It warns that neither is conclusive on its own. Blends can complicate burn observations, and damaged fibres can lose useful features.
For the opening bamboo-labelled sample, cellulosic burn behaviour would leave several possibilities. Cotton, regenerated cellulose and a mixture of cellulosic fibres still require further work. A familiar smell or ash appearance cannot establish the plant source or a percentage.
Use burning only within a suitable laboratory procedure. Do not turn it into a direct sniff test of unknown finishes or coatings. Some tests consume the material. Reserve enough for the later work needed to resolve the question.
Microscopy adds shape and structure
A microscope can reveal features that a bulk chemical result does not. Longitudinal views show the fibre along its length. Cross-sections show its shape across the width. These views answer related but different questions.
Cotton often has a ribbon-like appearance with twists. The Canadian Conservation Institute notes a change after mercerization. The twists can decrease, and the fibre can become more cylindrical. This is a useful reminder that processing can change a familiar identifying feature.
Compare the unknown with known reference material under the same viewing conditions. Examine more than one fibre and more than one location. One clear image may be useful. It does not prove that all fibres in the textile have the same identity.
A smooth fibre is not automatically one particular synthetic. Nor does a round cross-section uniquely name a polymer. Use these features to guide the next test. Several candidate materials may share the same shape.
Keep images tied to their sampling locations. A dark fibre might appear only on the reverse face. That may explain why tests on the two faces gave different results.
What FTIR contributes
Fourier transform infrared spectroscopy, or FTIR, gives evidence about chemical structure. It measures how a material interacts with infrared radiation. NIST describes its use to examine molecular vibrations. These give information about the material's structure.
In fibre work, the analyst compares the measured pattern with suitable reference spectra. The position and relative strength of several bands matter. One visible peak is rarely enough to support a specific identification.
The archived FBI fibre guide explains how to use reference spectra. It stresses that samples need to be prepared in the same way. The guide also notes a limit in separating rayon from lyocell by infrared spectra alone. This is dated guidance, not a substitute for a current laboratory method.
A library search can suggest possible matches. The analyst must still check the spectrum. Does the proposed match make chemical sense? A high software match score does not measure fibre purity. It is also not proof that every part of a fabric contains the same polymer.
The sampling mode can change the answer
An FTIR report should state how the spectrum was obtained. Common modes include transmission, reflection and attenuated total reflectance, known as ATR. They do not examine a sample in exactly the same way.
Thermo Fisher's explanation of ATR describes its value for surface layers and coatings. This can help with a coated textile. Still, check which layer the result describes. It may mainly show the coating, rather than the fibre beneath it.
Consider a fictional fabric with a polyester ground and a different polymer coating. A surface reading may be dominated by the coating. Calling that result the identity of the ground yarn would go beyond the sampled material.
Separate and examine the relevant parts where the method allows it. If they cannot be split, report the limit. Software may list one match first. That alone cannot turn a mixed signal into proof of a single fibre.
The same care applies to a core-and-cover yarn. Sampling only the outside can miss the core. The physical construction needs to guide the locations chosen for chemical analysis.
Use a second test to resolve a specific uncertainty
More tests do not always mean better evidence. Choose a second test that can tell the remaining alternatives apart. Repeating the same weak conclusion in a new format adds little.
A microscope might show two fibre groups while a bulk spectrum mainly shows one. Examine each group on its own to resolve this. Another sample may give clear evidence of cellulose. Its fibre form may still be unclear. Further optical work may help. When identity is clear but the percentage is in doubt, the task changes. Another qualitative result will not resolve the mass share.
ISO/TR 11827:2012 describes natural and manufactured fibre identification. It is intended to work with quantitative blend analysis where needed. ISO lists it as published, with a revision under development. Check the edition and status when specifying the work.
Solubility and other chemical tests can be part of a valid scheme. Their meaning depends on the actual materials and full method. An informal table of acid strengths is not enough to select a test for an unknown blend. Modified fibres, finishes and other parts may change the result.
Cellulose identity does not prove bamboo origin
Bamboo can be a feedstock for regenerated cellulose. That does not make every cellulosic result evidence of bamboo origin.
The US Federal Trade Commission gives guidance on bamboo claims. It distinguishes actual bamboo fibre from rayon made with bamboo feedstock. A test of the finished rayon alone does not prove which plant supplied the pulp. These are US labelling rules; other markets need their own review.
For laboratory planning, separate two tasks. First, establish the generic fibre identity. Then examine the evidence for the claimed feedstock. Do not ask a generic cellulose spectrum to prove both.
The same logic applies to performance claims. A rayon identity does not prove antimicrobial action. Nor does it establish skin benefits or environmental performance. Those claims need evidence about the claimed property and the actual product.
In the opening example, the appropriate next step depends on the purchase brief. The buyer may need a report on generic fibre content. Proof of bamboo feedstock may also be required. The laboratory request should say which question is being asked.
Brand authentication needs its own evidence
A brand can cover specific fibre grades and production routes. Finding the expected polymer may support a brand claim. It does not normally prove who supplied the fibre.
There are specialised systems designed for this purpose. Lenzing describes a fibre-identification technology for its own products. The company says it can verify their presence in fabrics and garments. It combines this with other traceability and licensing services. This system is specific to the supplier. It does not mean that any cellulose test can identify a Lenzing product.
When a buyer needs a branded fibre, check how to prove the source of that exact grade. Keep its result separate from the general composition report. A cellulose test and a valid brand check answer different questions. Both may be needed for the same sample.
A generic polymer spectrum cannot serve as proof of recycled content. Evidence must address the claimed source and how the material was made. A validated method designed for that specific claim may also be used.
Quantitative analysis begins after identity is clear
Once the fibre types are established, choose a method suited to the mixture. Some textiles allow physical separation of their components. Others need chemical or microscopic analysis under a validated procedure.
ISO 1833-1:2020 sets out general principles for chemical analysis. It concerns the quantities of fibres in mixtures. The appropriate part of the series matters because different mixtures need different procedures. Its public page identifies a revision in progress. A report should not silently substitute that draft for the published edition.
The mass basis must be clear. Does the result refer to dry fibre or conditioned material? Has nonfibrous matter been removed? Have the required corrections been applied? The label “composition” alone does not answer these questions.
The following examples show why the basis matters. These are invented mass balances. They do not give a fibre-dissolving recipe or replace a published method. All separations are assumed complete and selective unless stated otherwise.
Example: finish mass changes the denominator
An imagined dry specimen weighs 1.000 g before a validated pretreatment. The treatment removes 0.040 g of nonfibrous matter, leaving 0.960 g of dry fibre. Further analysis establishes 0.384 g of fibre B:
Fibre B share of dry fibre = 0.384 ÷ 0.960 × 100 = 40.0%
Dividing by the untreated mass instead gives:
0.384 ÷ 1.000 × 100 = 38.4%
The difference is 1.6 percentage points. No fibre identity changed; the denominator did.
This does not mean every finish can be removed by the same treatment. The method must establish what was removed and whether the fibres were affected. The example only shows why the report must identify its reference mass.
Example: a residue correction is not optional rounding
Suppose another fictional method starts with 0.500 g of prepared dry fibre and yields 0.190 g of residue. For this example only, assume a validated correction factor of 1.02 applies to that residue:
Corrected residue mass = 0.190 × 1.02 = 0.1938 g
Corrected fibre share = 0.1938 ÷ 0.500 × 100 = 38.76%
Without the correction, the result would be 38.00%. The correction accounts for the 0.76 percentage-point difference. Choosing more decimal places would not create it.
The factor is illustrative. Do not transfer it to a real fibre pair. Use the value and conditions required by the method. Keep enough calculation detail for another analyst to check the result.
Repeats reveal variation, not just a better average
Suppose three separately prepared portions of a fictional blend give 39.2%, 40.0% and 40.8% of fibre B on the same basis. Their mean is 40.0%, and the range is 1.6 percentage points. Keep those individual results when assessing the method and sample.
The mean alone hides the spread. It does not establish the uncertainty of the result. Nor does it prove that every part of the roll has that blend. How the sample is prepared and measured can affect the result. So can real changes within the material.
A result close to an agreed limit needs care. Use the laboratory's validated rules for reporting and decisions. More decimal places do not make a result more certain. A repeat may give a preferred result. It must not erase an earlier result without a sound technical reason on record. Check that the portions were alike and that the method worked as intended.
Counting fibres does not directly give their mass share
A microscope count may describe the number of fibres in the examined field. Fibre composition is often required on a mass basis. These are not automatically equal.
Take an invented set of equal-length fibre pieces. There are 80 pieces of fibre A at 1 dtex each and 20 pieces of fibre B at 3 dtex each. Their masses are proportional to number multiplied by linear density:
A mass units = 80 × 1 = 80
B mass units = 20 × 3 = 60
A mass share = 80 ÷ (80 + 60) × 100 = 57.14%
A makes up 80% of the count but only 57.14% of the mass. B makes up 20% of the count and 42.86% of the mass.
Real methods that use a microscope to measure blend content must account for sample choice and fibre size. This simple example is not such a method. It shows why a raw count should not be relabelled as a mass percentage.
ApparelDiary's yarn count guide explains the mass-per-length units. Its moisture regain guide helps distinguish moisture from dry material in weighing calculations.
A minor component can be missed
Not finding a fibre in a small test does not prove that it is absent from the textile. What was sampled and what the test can detect both matter.
Consider a deliberately simple model. Assume a minor fibre makes up 2% of the fibre count and every selected fibre is an independent random pick. The chance of missing it in all 50 picks is:
Chance of no detection = (1 − 0.02)⁵⁰ ≈ 36.4%
At 150 independent picks, the same model gives about 4.83%. These figures are not recommended sample sizes. They show that a small examination can miss a minor fibre. This can happen even with ideal random sampling.
Real fabrics can be less simple. A component may occur only in certain yarns, stripes, faces or cores. Repeated picks from one yarn do not cover the whole textile independently.
A report should state what was examined and the relevant limit. “Not detected in the examined fibres” may be the correct finding. Do not claim complete absence unless the method supports it.
NIR screening needs a validated classification model
Near-infrared spectroscopy, or NIR, can support rapid textile sorting. Its model commonly links the measured spectra to known materials.
NIST's NIR-SORT project provides reference data for developing and evaluating these systems. It includes known-provenance and post-consumer textiles. NIST identifies a major challenge here: access to suitable reference data. Without it, comparing how well systems classify textiles is difficult.
For a mill or sorting operation, test the system on the materials it will actually encounter. Include the relevant blends, colours, finishes and constructions in the validation set. A model trained on simple materials may struggle with complex textiles. It needs evidence for those uses too.
Keep confidence in the assigned class separate from blend content. A software score of 95 for a polyester class does not mean the sample contains 95% polyester. To report a mass percentage, validate the system for that measurement. It needs suitable calibration and a stated uncertainty.
Route doubtful or unfamiliar samples for further analysis. The system needs a way to flag weak evidence. It should not force every spectrum into the closest available class.
Resolve conflicting results by checking what each test saw
Suppose a fictional fabric gives a polyester spectrum on the face. A microscope shows different fibres in the separated back yarns. The results may both be correct. They concern different components.
Before repeating every test, compare the sample locations and preparation. Were the same yarns examined? Was a coating included? Was the spectrum collected from a blend or an isolated fibre? Did a cleaning step remove material that another test still included?
Then choose the next examination to resolve the difference. A second face scan will not necessarily explain the back yarn. Take a fresh sample from the same identified location. It may tell more than another cutting with an unknown position in the roll.
The report should preserve unresolved findings. The evidence may support only a broad class. Use that class and state what further work could narrow it. Editing must not turn doubt into a confident claim about brand or origin.
Specify the method and report the actual conclusion
Method status is part of technical accuracy. ASTM D276-12 is marked withdrawn in 2021 on its edition page. Older explanatory text remains on the site. A general search snippet can therefore mislead readers about its status.
For a new request, agree on a suitable current method and obtain its full procedure. Public summaries help with that choice. They leave out some steps, corrections and rules for reports. These details appear in the full method.
A useful report names the submitted material and the parts sampled. It records the methods, editions and results on their stated basis. Keep observed evidence distinct from the analyst's conclusion. State any unresolved fibre identity, blend share or origin claim.
This makes the result useful beyond the laboratory. A buyer can see whether the composition brief was checked. A technologist can see which component caused a process concern. A product team can see which claims the material test supports. It can then seek any missing evidence about origin.
Discussion0
Loading the discussion…