🍎 Analysis of milk and liquid milk products by mid-infrared spectrometry (MIR/FTIR)
In short
After pretreatment and homogenisation the milk sample is passed through the flow cell of a spectrometer, which measures the absorption of mid-infrared radiation at selected wavelengths. The test is performed according to PN-ISO 9622:2015-09; Character of the method: a secondary instrumental method — the result arises from a calibration model built on samples determined by reference methods, not from the spectrum itself. The procedure comprises 10 steps; it is used for: Assessment of the composition of raw milk on collection and in settling accounts with suppliers, Milk recording of cows (ICAR) — determination of fat, protein, lactose and urea by MIR as a routine method, Control of the composition of processed milk, cream and whey in a dairy plant.
At a glance
- Standard: PN-ISO 9622:2015-09
- Category: Food chemistry
- Procedure steps: 10
- Character of the method: a secondary instrumental method — the result arises from a calibration model built on samples determined by reference methods, not from the spectrum itself
- Matrices covered by the guidelines: raw milk, processed milk, cream, whey; the guidelines are written for cow’s milk and are applicable to the milk of other species after a separate calibration
- Limitation of use: only products of lower viscosity, capable of being pumped through the flow system of the analyser, and analytes not causing optical saturation at the wavelengths used
Overview
The analysis of milk in the mid-infrared is the basis of the mass assessment of the raw material in dairies and in milk-recording laboratories: a single measurement takes a few tens of seconds and gives several parameters of the composition at once. PN-ISO 9622:2015-09 implements identically (IDT) the international standard ISO 9622:2013, drawn up jointly with the International Dairy Federation as ISO 9622|IDF 141. The standard is not a recipe for a single determination — it is a set of GUIDELINES for the quantitative determination of the composition of milk and liquid milk products (raw milk, processed milk, cream, whey) by measuring the absorption of radiation in the mid-infrared range.
The heart of its application: it is a SECONDARY instrumental method. The standard itself says that the spectral data are converted into estimates of the concentrations of the constituents by calibration models built on samples representative of the population examined. The ICAR guidelines for milk-recording laboratories call the matter by its name: "reference methods" are the methods used for calibrating routine methods and should be internationally standardised methods (ISO, IDF, AOAC). For fat the reference method is the Röse-Gottlieb gravimetry (ISO 1211 | IDF 1), for total protein titration after Kjeldahl digestion (ISO 8968 | IDF 20, parts 1 and 3), for lactose HPLC (ISO 22662 | IDF 198); as secondary reference methods there are admitted among others the Gerber butyrometry (ISO 19662 | IDF 238) for fat, the enzymatic method (ISO 5765 | IDF 79) and the differential pH method (ISO 26462 | IDF 214) for lactose. MIR spectrometry itself appears in the ICAR table of standardised routine methods — with a reference to ISO 9622 | IDF 141 and AOAC 972.16.
A caution about editions and about the title. The edition in force, PN-ISO 9622:2015-09 (implementing ISO 9622:2013), bears the title „Mleko i płynne przetwory mleczne — Wytyczne do stosowania spektrometrii w zakresie średniej podczerwieni" [Milk and liquid milk products — Guidelines for the application of mid-infrared spectrometry] and covers a wide range of parameters and matrices. The title speaking of the determination of milk fat, protein and lactose belongs to the PREVIOUS edition — PN-ISO 9622:2006 (implementing ISO 9622:1999, scope: analysers for the determination of fat, protein and lactose in farm milk), which was withdrawn on 24.09.2015 and superseded by the 2015 edition. The PKN catalogue card for PN-ISO 9622:2015-09 has neither a "Date of withdrawal" nor a "Superseded by" field — this is the edition in force (33 pages in the 2006 edition, 20 pages in 2015; technical body: KT 35 for Milk and Milk Products).
What this description does not contain: we have not seen the full text of the standard. The scope, the principle, the construction of the instrument and the factors influencing the measurement (clauses 1–6) are given from the open preview of ISO 9622:2013 covering pages 1–5; the clauses on calibration, on the determination itself, on the control of stability and on precision (7–12) are not covered by the preview — the limit values of repeatability, slope and bias given below come from the ICAR guidelines, which derive some of them expressly from ISO 9622 | IDF 141 and mark others as indicative in the absence of values in the standards. The analytical wavelengths, the construction of the calibration models and the limits of quantification for a particular analyser we do not guess — they are taken from the text of the standard and from the manufacturer’s documentation.
Method principle
After pretreatment and — if it is required — homogenisation, the sample is measured with an infrared spectrometer which records the amount of radiation absorbed in the transmission mode at defined wavelengths in the mid-infrared range. The spectral data are converted into estimates of the concentrations of constituents or of other physicochemical parameters by means of calibration models developed on samples representative of the population examined; for some parameters (e.g. the quantity equivalent to the freezing point, the freezing point equivalent) signals from other sensors of the instrument are additionally introduced into the model. The signal at the appropriate wavelengths is obtained either from an interferogram subjected to a Fourier transform (FTIR) or by means of optical filters; instruments and calibration models differ in the number of wavelengths used to estimate a given parameter. The standard distinguishes spectral calibration (a model combining absorbance signals from many — more than two — wavelengths or signals from other sensors, optimised mathematically) and slope-and-intercept calibration (the coefficients of a straight-line regression of the instrument readings against the results of the reference methods). With classical calibrations based on the absorbance at fixed wavelengths (MLR models) it is necessary to apply intercorrections, which compensate the influence of changes in the concentration of the remaining constituents — the intercorrection coefficients are specific to each wavelength and each type of instrument.
Applications
- Assessment of the composition of raw milk on collection and in settling accounts with suppliers
- Milk recording of cows (ICAR) — determination of fat, protein, lactose and urea by MIR as a routine method
- Control of the composition of processed milk, cream and whey in a dairy plant
- Analysis of the milk of other species (goat, sheep, buffalo) and of the liquid products derived from it — provided there is a separate calibration and separate control procedures for each application
- Internal quality control of analysers: repeatability, zero stability, linearity, carry-over, efficiency of the homogeniser
Key parameters
| Parameter | Value |
|---|---|
| Character of the method | a secondary instrumental method — the result arises from a calibration model built on samples determined by reference methods, not from the spectrum itself |
| Matrices covered by the guidelines | raw milk, processed milk, cream, whey; the guidelines are written for cow’s milk and are applicable to the milk of other species after a separate calibration |
| Limitation of use | only products of lower viscosity, capable of being pumped through the flow system of the analyser, and analytes not causing optical saturation at the wavelengths used |
| Reference methods for calibration (ICAR list) | fat — ISO 1211 | IDF 1 (Röse-Gottlieb gravimetry); total protein — ISO 8968 | IDF 20 parts 1 and 3 (Kjeldahl); lactose — ISO 22662 | IDF 198 (HPLC) |
| Check of the homogeniser | the difference between five replicate fat readings for unhomogenised and homogenised milk must not exceed 0.04 % for milk with a fat content of 4.0 % (for other contents: fat content × 0.01) |
| Carry-over | determined from a series of 20 samples in the sequence water, water, milk, milk…; the calculated values Ew and EM must be smaller than ±1 % |
| Linearity | at least 10 solutions of known concentration over the typical range of the constituent, measured in triplicate; the curvature index r should be smaller than 2 % |
| Control limits according to ICAR (fat, protein, lactose) | repeatability sr 0.014 % units (the limit from ISO 9622 | IDF 141), intercorrections ±0.02 % units; indicatively: mean bias ±0.02 % units, slope 1.00 ±0.03, daily stability ±0.05 % units, zeroing ±0.03 % units |
Standard
- Standard number
- PN-ISO 9622:2015-09
- Title (PL)
- Mleko i płynne przetwory mleczne — Wytyczne do stosowania spektrometrii w zakresie średniej podczerwieni
- Title (EN)
- Milk and liquid milk products — Guidelines for the application of mid-infrared spectrometry
Step-by-step procedure
-
Preparing the instrument
Start the analyser in accordance with the manufacturer’s instructions. Check the state of the drier of the optical compartment — replace the silica gel before it begins to change colour.
-
Control of repeatability
Examine a homogeneous, representative sample at least 12 times in succession; discard the first two results in order to eliminate the influence of carry-over. The calculated repeatability is to lie within the limits proper to the parameter and the matrix. According to ICAR the assessment of repeatability is carried out at the start-up of every instrument on 10 replicates of one control sample.
-
Control of the zero
Analyse a zero sample (water or zero solution) periodically. The drift is to be small and random in sign, so that the cumulative drift is negligible; a plot of the drift against time is helpful. In instruments with factory auto-zeroing, review the automatic corrections.
-
Check of the homogeniser
Carry out two successive analyses: of unhomogenised whole milk and of the same sample after it has passed through the homogeniser of the instrument. The difference between five replicate fat readings must not exceed 0.04 % for milk with a fat content of 4.0 %. Note: an instrument with a completely inoperative homogeniser gives a very small difference — the result of the test is sometimes misleading.
-
Check of linearity
Prepare at least 10 solutions of known concentration covering the typical range of the constituent (serial dilutions from the stock solution). Examine each sample in triplicate, average, determine the linear regression y = bx + a, calculate the residuals and plot them against the expected values. The curvature index r is to be smaller than 2 %.
-
Check of carry-over
Examine 20 successive samples of water and whole milk in the sequence: water, water, milk, milk, water, water… Calculate the carry-over water→milk (Ew) and milk→water (EM); both values must be smaller than ±1 %. To assess the whole system (with the sampler) the samples are presented from 20 separate vials; for the flow system alone — manually, wiping the pipette between cycles.
-
Calibration of the instrument
Build or check the calibration model on samples representative of the population examined, with reference values obtained by standardised methods (fat — ISO 1211 | IDF 1, protein — ISO 8968 | IDF 20, lactose — ISO 22662 | IDF 198). With MLR models set the intercorrections compensating the influence of the remaining constituents. The detailed procedure of calibration and of checking the slope and the intercept is given in clause 7 of the standard — it has to be read in the original.
-
Measurement of the samples
Prepare the samples and bring them to the state required by the manufacturer of the instrument, then present them to the flow system. The application is limited to products of lower viscosity, capable of being pumped through the system, and to analytes not causing optical saturation.
-
Ongoing control of stability
In the course of the work analyse control (pilot) samples to assess the daily and short-term stability; according to the ICAR recommendations the frequency is about 3 times an hour for short-term stability and 4 times a day for zeroing, while the slope and the bias are checked monthly and weekly respectively.
-
Response to exceedances and test report
On an exceedance of the control limits, readjust the settings of the instrument (clause 10.5 of the standard) before further measurements. In the test report give the method together with a reference to PN-ISO 9622:2015-09 and the information that the result comes from an instrumental method calibrated by reference methods.
Required equipment and apparatus
| Equipment | Example | Indicative price |
|---|---|---|
| Mid-infrared milk analyser | an instrument with an interferometer and a Fourier transform (FTIR) or with optical filters, with a flow cell | — |
| Homogeniser built into the analyser | it breaks up the fat globules before the measurement; its efficiency is checked by a separate test | — |
| Flow system with pump and feeding pipette | the source of carry-over — checked together with the automatic sampler | — |
| Automatic sampler | equipment typical of flow analysers; ICAR requires the sample to be mixed immediately before the measurement, but the sources do not describe the way of mixing in the sampler | — |
| Drier of the optical compartment (silica gel) | replaced before it begins to change colour, and not less often than at the interval given by the manufacturer — moisture in the compartment changes the optical zero and the calibration | — |
| Set for preparing control samples and linearity solutions | a balance for weighing out fractions, a water bath for thermostatting the samples | — |
Reagents, media and consumables
| Reagent | CAS | Details |
|---|---|---|
| Zero sample (water or zero solution) | — | analysed periodically in the course of the work to control the stability of the zero; the drift should be small and random in direction |
| Homogenised cream with a fat content of 8 % | — | for checking the linearity of the fat channel — diluted with skimmed milk or with the zero solution |
| Retentate from the ultrafiltration of skimmed milk | — | for the linearity of the protein channel, diluted with ultrafiltrate; stock solution with a protein content of about 5.5 %. Alternatively whey protein concentrate, sodium caseinate, calcium propionate [TO BE CHECKED IN THE ICAR GUIDELINES: calcium propionate contains no protein, so as a solution for the linearity of the protein channel it makes no sense — possibly the source has calcium caseinate; we do not decide, because we do not have the guidelines to hand], skimmed milk powder or concentrated skimmed milk diluted with distilled water |
| Lactose monohydrate solution 60 g/l | — | for the linearity of the lactose channel — diluted with water or with a solution of milk mineral salts |
| Control (pilot) milk samples | — | for daily and short-term control of stability; according to ICAR prepared under a quality assurance regime (control of homogeneity and stability), with assigned values linked to standards or reference materials |
Health and safety (OHS)
- Raw milk is potentially infectious material (among others pathogenic microorganisms) — work in gloves, do not pipette by mouth, wash hands after work.
- Milk samples are sometimes preserved with preservatives — before work check the safety data sheet of the preservative used and comply with it.
- The analyser contains a source of infrared radiation and live parts — do not open the optical compartment while it is working, leave servicing to the manufacturer.
- The homogeniser works at high pressure — do not open the flow system while the instrument is running.
- Collect the waste after the analysis (milk, rinsing water) in accordance with the plant’s waste management procedure; do not pour solutions with added preservatives into the drain.
Frequently asked questions
Which standard describes this test?
The test is performed according to PN-ISO 9622:2015-09 — “Milk and liquid milk products — Guidelines for the application of mid-infrared spectrometry”.
How does this method work?
After pretreatment and — if it is required — homogenisation, the sample is measured with an infrared spectrometer which records the amount of radiation absorbed in the transmission mode at defined wavelengths in the mid-infrared range. The spectral data are converted into estimates of the concentrations of constituents or of other physicochemical parameters by means of calibration models developed on samples representative of the population examined; for some parameters (e.
What is the measuring range and accuracy?
Character of the method: a secondary instrumental method — the result arises from a calibration model built on samples determined by reference methods, not from the spectrum itself; Matrices covered by the guidelines: raw milk, processed milk, cream, whey; the guidelines are written for cow’s milk and are applicable to the milk of other species after a separate calibration; Limitation of use: only products of lower viscosity, capable of being pumped through the flow system of the analyser, and analytes not causing optical saturation at the wavelengths used; Reference methods for calibration (ICAR list): fat — ISO 1211 | IDF 1 (Röse-Gottlieb gravimetry); total protein — ISO 8968 | IDF 20 parts 1 and 3 (Kjeldahl); lactose — ISO 22662 | IDF 198 (HPLC).
How long does the test take?
The procedure comprises 10 steps. The standard does not give a duration for every stage — the laboratory's own procedure decides.
What equipment is required?
Mid-infrared milk analyser, Homogeniser built into the analyser, Flow system with pump and feeding pipette, Automatic sampler, Drier of the optical compartment (silica gel), Set for preparing control samples and linearity solutions.
Where is this test used?
Assessment of the composition of raw milk on collection and in settling accounts with suppliers; Milk recording of cows (ICAR) — determination of fat, protein, lactose and urea by MIR as a routine method; Control of the composition of processed milk, cream and whey in a dairy plant; Analysis of the milk of other species (goat, sheep, buffalo) and of the liquid products derived from it — provided there is a separate calibration and separate control procedures for each application; Internal quality control of analysers: repeatability, zero stability, linearity, carry-over, efficiency of the homogeniser.
What safety precautions apply?
Raw milk is potentially infectious material (among others pathogenic microorganisms) — work in gloves, do not pipette by mouth, wash hands after work.; Milk samples are sometimes preserved with preservatives — before work check the safety data sheet of the preservative used and comply with it.; The analyser contains a source of infrared radiation and live parts — do not open the optical compartment while it is working, leave servicing to the manufacturer.; The homogeniser works at high pressure — do not open the flow system while the instrument is running.; Collect the waste after the analysis (milk, rinsing water) in accordance with the plant’s waste management procedure; do not pour solutions with added preservatives into the drain.
Which laboratory can perform this test?
The test is performed by laboratories accredited to ISO/IEC 17025. On LabCoda you can find them by the standard number PN-ISO 9622.