process FTIR
How to Evaluate Process FTIR for Real-Time Chemical Measurement
Turn a Process FTIR Question into a Supplier Test Brief
A useful supplier discussion starts with a process question and ends with a defined test request. For process FTIR, prepare a supplier test brief that lets a technical team map your stream, measurement conditions, calibration evidence, and installation needs to a proposed system. This is more useful than comparing spectrometer specifications in isolation. In-situ FTIR can continuously capture spectral data as chemistry evolves, supporting real-time reaction monitoring, process optimization, and automated process control. Turning that spectral stream into a concentration estimate still requires a reference basis, a suitable multivariate or chemometric model, and evidence that the model addresses the intended process conditions.
Use the brief to make each response traceable. For every question, record the process requirement, the supplier’s response, and the evidence or test that would support the response. Mark whether the information is already available, must be measured on representative material, or remains unresolved. The resulting document gives both sides a shared handoff without treating a general FTIR capability as proof of suitability for an unspecified stream.
-
State the measurement request in process terms. Name the analytes or properties of interest, the required concentration range and its extremes, the composition of the material, and the result that the process team needs to use. Include the intended sampling frequency or response time, the required accuracy, and whether the measurement is intended for a reaction, continuous process, recirculation loop, pilot rig, or bench setup. This defines what the supplier must demonstrate rather than leaving the discussion at “real-time analysis.”
-
Ask how the spectrum becomes an estimate. Request a description of the path from spectral features to concentration or other reported value. The brief should identify the reference analytical procedure, the samples and reference values used for model development, the multivariate or chemometric approach, and the way results will be assessed. PLS regression can handle strongly correlated FT-IR variables, but the reference procedure remains part of the model’s foundation. Record its reproducibility and the expected effect of reference-data variation instead of treating the instrument signal as a direct concentration result. Multivariate data analysis can reduce spectral-data complexity and support rapid processing for real-time monitoring.
-
Describe the process change envelope. List the temperature, humidity, pressure, composition, concentration range, water content, solids, bubbles, flow conditions, and other changes that the measurement may encounter. Then ask which of those conditions are represented in the calibration material and which require additional samples or testing. Include changes in feedstock, product specifications, measurement conditions, or instrument configuration in the supplier’s response. Paeonia Innovations offers ongoing chemometric-model support, including revalidation as feedstock, temperature, or product specifications change; the brief should clarify what revalidation would mean for the particular application.
-
Specify how material reaches the optics. Describe the proposed probe, flow cell, bypass, or side-stream arrangement, along with pressure, temperature, flow, residence time, optical access, cleaning expectations, and the presence of water, solids, or bubbles. The material presented to the measurement must represent the process question being asked. Installation details therefore belong in the test brief, not as a later procurement detail. If a flow or sampling condition affects the result, identify the measurement and record that would resolve the uncertainty.
-
Define the evidence handoff. Identify representative samples for calibration and separate samples for challenge. Validate a multivariate model with independent test data and compare its results with predetermined performance criteria. State which concentration extremes and process changes the test must cover. A calibration set must span the expected concentration range, including the extremes, and the reference method’s reproducibility places a boundary on the model output. If the proposed material does not represent the planned process, record that limitation and request the additional testing needed before relying on the result.
The supplier test brief can then be used to compare documented capabilities with the application. Paeonia Innovations offers the OrionIR® mid-IR spectrometer, providing one spectrum every second across 900–3500 cm⁻¹ through aqueous, strongly absorbing, and low-concentration streams. We build chemistry calibration models and integrate them into instrument software to provide real-time concentration readouts. We also offer to test our technology on your process samples before commitment. Those are specific capabilities to examine against the stream and evidence requirements; they do not by themselves establish suitability for a particular process.
For sampling arrangements, we designed the OrionIR® Lab version for applications such as reactor recirculation loops, pilot-rig side streams, and bench setups. The measurement principle of the OrionIR® High-Pressure Version is unchanged from the Lab version, allowing chemometric models built during evaluation to transfer to a pressurized installation. Treat that documented capability as a question for the planned installation: identify the pressure conditions, the material used during evaluation, and the evidence needed to confirm that the transfer applies to the reader’s setup. Paeonia Innovations has also field-tested the OrionIR® under industrial conditions in North America, South America, Europe, and Asia, including a repeatability test in which the instrument held spectral stability for more than 100 days. These owner-attested facts provide context for a supplier discussion, not an application-specific performance result.
A message requesting review can therefore include five attachments or sections: the process question, the sample and concentration description, the reference-method information, the operating and installation envelope, and the independent-test request. Ask the supplier to identify which points are supported by existing information, which require process-sample testing, and which remain outside the documented capability. This makes the next conversation about the reader’s actual stream rather than about a generic instrument category.
The decision boundary should remain explicit. General FTIR principles explain how spectral data, reference values, and models can be connected. Product information describes what Paeonia Innovations offers. Neither category establishes regulatory acceptance, a safety classification, or performance for an unspecified process stream. A supplier feasibility discussion is warranted when the unresolved points have been listed and the proposed samples, reference procedure, performance criteria, and installation conditions are clear enough to review.
What process FTIR measures continuously
The process measurement captures changing chemical information through a stream or reaction environment by recording spectra as the chemistry evolves. That measurement mode differs from laboratory-only analysis, where an operator collects samples and obtains discrete snapshots at selected times. In-situ FTIR continuously captures spectral data, while offline sampling provides observations at predetermined time points.
The distinction matters because a continuous spectrum is not automatically a continuous concentration result. The spectrum is the measured input. Concentration estimates require a relationship between spectral features and reference values, usually represented by a calibration model. The usefulness of the final result depends on whether the material, concentration range, and measurement conditions represented during calibration match the process you want to monitor.
A process measurement also has a physical location. You may measure in a vessel, through a recirculation loop, in a bypass, or on a side stream. Each arrangement presents a particular portion of the process to the optical measurement. A laboratory spectrum from a prepared sample can help characterize chemistry, but it does not by itself demonstrate that an installed arrangement will receive representative material at the required frequency.
Use continuous measurement when the process question concerns how composition changes over time, not simply whether a sample met a laboratory specification at one collection point. Define the concentration estimate, response timing, and process decision that the result must support before selecting the instrument configuration. If the decision depends on a short-lived change, include that event in the test design rather than relying only on stable operating samples.
Overview
Process FTIR suitability is an evidence-chain decision. The direct answer is conditional: investigate the approach when the spectrum contains information related to the required measurement, representative process material can support calibration, independent data can test the model, and the installation can present that material consistently.
The evaluation can be organized as a short sequence:
- Define the output. State which analytes or chemical properties you need to estimate, the concentration range, the update frequency, and the decision that will use the result.
- Characterize the input. Describe the process stream, including composition, expected changes, temperature, pressure, water content, solids, bubbles, and the proposed optical path or sampling arrangement.
- Build the relationship. Use reference standards or process samples with reference values, spectral data, and a multivariate or chemometric model to connect the spectrum to an estimated concentration.
- Challenge the relationship. Validate the multivariate model with independent test data and compare the results with performance criteria established in advance.
- Review the installation. Determine whether the instrument, interface, flow path, software, and support model fit the evidence requirements.
No single instrument specification answers all five questions. Spectral coverage may determine what information is available, while the calibration and installation determine whether that information becomes a usable process result. Keep those questions separate so that a strong result in one area does not substitute for missing evidence in another.
This framework also helps you assign the next task. A missing analyte range calls for better process characterization; an uncertain reference method calls for analytical review; inconsistent material presentation calls for sampling or installation work; and weak independent results call for model investigation. Each gap has a different remedy.
From spectral data to concentration estimates
A concentration estimate comes from a model linking measured spectra to reference values; the spectrum itself is not the finished concentration result. The workflow begins with material for which the target analyte or property has been measured by a reference procedure. The corresponding FTIR spectra provide the model inputs, and the reference results provide the values the model is trained to estimate.
A practical workflow contains these stages:
- Collect representative material. Include the process states and composition changes that the installed measurement is expected to encounter.
- Acquire spectra. Record the spectral response through the selected optical arrangement under the conditions relevant to the application.
- Assign reference values. Pair each spectrum with a reference-method result for the target analyte or property.
- Fit the model. Multivariate or chemometric methods use information distributed across the spectral variables rather than relying on a single isolated feature.
- Calculate estimates. The model converts a new spectrum into an estimated concentration or other defined process value.
- Test the estimates. Compare predictions from independent samples with predetermined performance criteria.
Partial least-squares regression can handle numerous, strongly correlated variables such as FT-IR spectral variables, but a reliable reference method is necessary for that model. Multivariate data analysis can reduce spectral-data complexity and support rapid processing for real-time monitoring. Those computational capabilities do not remove the need to define the process domain or test the model with material outside the calibration set.
The practical question is therefore not whether a model can be fitted. A model can be fitted to many datasets. The useful question is whether the fitted relationship remains sufficiently informative when a new process spectrum arrives from the intended stream, interface, and operating range. Record the path from sample to estimate so that a later discrepancy can be traced to the reference result, spectrum, model, or physical measurement arrangement.
Why the reference method sets the analytical boundary
The reference method sets a practical boundary for a spectroscopic model because the model learns from the reference values paired with the spectra. Reference-method noise can affect the accuracy and precision of spectroscopic predictions. A 2002 study reported acceptable predictions when the noise in the reference data was not too large, which does not make poor reference data acceptable for every application.
Our calibration practice follows the same constraint: our predicted values cannot be more accurate than the reference method used for training, and the reference method’s reproducibility sets the model’s accuracy floor. This is a business-specific calibration principle, not a universal performance promise for every process or reference procedure.
Before asking a supplier to build a model, document the reference procedure in enough detail to assess its role in the result:
- What analyte or property does it measure?
- Which sampling and preparation steps can introduce variation?
- What repeatability and reproducibility are expected across operators, instruments, laboratories, or days?
- Does the reference result describe the same material and time point represented by the spectrum?
- Is the reference uncertainty compatible with the process decision you want to make?
A narrow reference range can also hide the model’s behavior at the edges of the process domain. Reference results should cover the concentrations and conditions that matter to the intended measurement. If the reference procedure cannot distinguish the process states that your decision separates, improving the spectrometer alone will not resolve that boundary.
Keep reference results and spectral observations linked by sample identity, collection time, process condition, and preparation record. That record does not improve the reference method itself, but it lets you determine whether a disagreement reflects a real process difference or a mismatch between the material used for the two measurements.
Define the process domain before judging fit
The process domain is the set of chemical and physical conditions represented by the calibration and expected during measurement. Define that domain before deciding whether a model is suitable. A calibration may need redevelopment when sample composition differs materially from the composition represented during model development.
Describe the domain in a written application brief:
- Target outputs: Name each analyte or property and state whether the result is qualitative, quantitative, or used for a threshold decision.
- Concentration range: Give the expected minimum, maximum, normal operating range, and meaningful extremes. Our calibration set must span the expected concentration range, including the extremes.
- Composition: List major components, expected feedstock changes, additives, by-products, impurities, and possible formulation changes.
- Physical conditions: Record temperature, pressure, water content, solids, viscosity where relevant to the interface, and the presence of bubbles or suspended material.
- Measurement setting: Describe the vessel, pipe, recirculation loop, bypass, side stream, or bench arrangement and the conditions under which samples will be collected.
- Required accuracy and timing: State the error boundary, update interval, and process action connected to the result.
One cited product study illustrates why the domain must be explicit: its two analytes were spiked across 21 samples at 2%–20% and 0%–14%, respectively, and the authors described redevelopment when composition differed from the represented formulation. Those ranges belong to that study; they are not default ranges for your process.
Mark which conditions are already represented by samples and which remain hypothetical. That distinction prevents a nominal operating range from being mistaken for demonstrated model coverage.

Treat optical access and sampling as part of the result
The sampling and optical arrangement determines which material reaches the measurement, so the interface belongs inside the analytical evaluation. A spectrum from a clean laboratory sample does not establish that a probe, flow cell, bypass, or side stream will present representative process material under operating conditions.
Define the physical arrangement before interpreting model results:
- Identify where material will be withdrawn or measured and how it returns to the process, if it returns at all.
- Record pressure, temperature, flow, residence time in the sampling path, and any conditioning steps.
- State how water, solids, suspended material, bubbles, fouling, and phase changes will be handled during the test.
- Specify the optical path length, window or cell arrangement, cleaning approach, and access needed for maintenance.
- Compare the material in the sampling path with the material represented by the reference samples.
- Establish how flow, sampling-system effects, and data gaps will be recorded.
The 2017 United States EPA Method 320 provides a concrete example of why flow conditions require attention. For an accurate mass flow meter, the method reports accuracy of ±1 percent of span; it gives a 5 percent flow-measurement uncertainty when 1 L/min is monitored with a meter calibrated over a 0–5 L/min range, and notes that recalibration at the specific operating flow may improve the result. Those figures describe the cited method and example, not every installation.
Use the example as a prompt for a supplier discussion: what flow range is measured, where is the meter calibrated, how are sampling effects evaluated, and how will the test show that the material reaching the optical path represents the process question? A model cannot correct for a material stream that never reaches the measurement in the required form.
Also define the relationship between the sampling point and the process decision. If the measurement is taken from a bypass or side stream, record how that location relates to the vessel or line being controlled. If conditioning changes temperature, pressure, phase, or residence time, include those changes in the representative test rather than treating the conditioned sample as identical to the source stream. The purpose is not to prescribe one installation, but to make the material path part of the evidence.
Validate the model with independent process evidence
Independent process evidence tests whether a calibration remains useful beyond the data used to build it. International pharmaceutical analytical validation guidance defines model validation as challenging a model with independent test data and comparing the results with predetermined performance criteria. That sequence should be visible in a feasibility review.
Separate the model-building dataset from the test dataset before reviewing results. Independent samples may come from a defined source that was not used for model development, provided the test design matches the question you need to answer. Do not choose acceptance criteria after seeing the predictions.
Set out the validation plan in advance:
- Identify which samples or process periods will remain outside model development.
- Define the concentration or property range that the test must cover.
- State the performance measures and decision thresholds.
- Record temperature, pressure, composition, sampling arrangement, and instrument conditions for each test observation.
- Compare the model output with the reference result and document exceptions.
- Decide whether the result supports the stated process decision, requires additional data, or stops the evaluation.
A passing result on independent samples answers one question: whether the tested model performed against the selected criteria for that test domain. It does not automatically establish performance after a major composition change, instrument change, installation change, or operating-condition change. Those conditions require their own transfer or validation questions.
Keep the test result tied to the conditions under which it was obtained. A single aggregate performance value can hide failures at concentration extremes, during composition changes, or in a particular sampling arrangement. Review the individual observations and the conditions attached to them before deciding whether the model answers the process question.

Separate calibration validity from model transfer
Calibration validity and model transfer are related but separate questions. Calibration validity asks whether the model describes the domain used to develop and test it. Model transfer asks whether that relationship remains usable after the instrument, environment, material, or measurement conditions change. An existing multivariate model may not remain valid when spectra are measured under different temperature or humidity conditions or on a separate instrument without adjustment.
Make transfer explicit in the test plan. Ask whether the following will change between development and intended use:
- spectrometer or optical configuration;
- temperature or humidity around the measurement;
- pressure and installation arrangement;
- sampling path, flow regime, or optical path;
- feedstock, formulation, product specification, or impurity profile;
- maintenance state or other condition that changes the measured spectrum.
A calibration may need redevelopment when the sample composition differs materially from the composition represented during model development. That boundary applies even if the new material looks similar to the earlier material at a high level. Define what similarity means through analytes, concentration range, matrix, and reference results.
Treat calibration transfer as a defined technical investigation rather than assuming that an existing model remains valid after an instrument or measurement condition changes. This does not establish transfer for another stream, instrument pair, or installation. Ask the supplier which transfer samples, adjustment method, and independent criteria will be used for your conditions. If those conditions cannot be represented, keep the conclusion bounded to the tested setup.
Separate transfer evidence from maintenance planning. A model may be tested successfully after one instrument change while remaining untested after a sampling-path change or a new product specification. List each change that could alter the measured spectrum, assign the corresponding test, and record the boundary beyond which redevelopment rather than adjustment is required.
Balance spectral coverage with process practicality
Spectral coverage should be matched to the chemical question and installation, rather than treated as a universal ranking of instruments. An instrument evaluation should consider the trade-off between wavelength coverage, cost, signal-to-noise ratio, and resolution. High-resolution instruments covering a full wavelength range may be relatively expensive for in-line process analysis, while specialized instruments focused on required wavelengths can reduce cost but typically provide lower signal-to-noise ratio and resolution.
Use the target measurement to frame the trade-off:
- Which spectral information is needed for the analytes or properties in the process brief?
- Does the selected range contain that information under the expected matrix and concentration conditions?
- What signal-to-noise ratio and resolution are required for the reference-method boundary and decision threshold?
- Does the installation need broad exploratory coverage, or is a focused range sufficient for a defined model?
- What response interval and data-processing path are required by the process decision?
Processing speed, however, does not establish chemical validity. A fast estimate from a poorly matched domain remains a poorly supported estimate.
Compare instruments against the same application brief. A narrower instrument may be a sensible fit for a defined measurement, while broader coverage may be useful during exploratory work. Let the required analytes, matrix, reference method, and validation plan determine which trade-off deserves further testing.
For an academic or engineering evaluation, preserve the distinction between exploratory breadth and routine measurement requirements. Broad data can help you investigate which spectral information is useful, while a focused configuration may be evaluated later against the defined model and installation. Neither choice removes the need to show that the selected range supports the required estimate under representative conditions.
Use documented supplier capabilities as fit questions
We document several OrionIR® capabilities that can be compared with your process brief. We offer the OrionIR® mid-IR spectrometer, providing one spectrum every second across 900–3500 cm⁻¹ through aqueous, strongly absorbing, and low-concentration streams. We build chemistry calibration models and integrate them into instrument software to provide real-time concentration readouts. We also offer to test our technology on your process samples before commitment.
Those statements identify questions for application fit; they do not decide fit for an unspecified stream. Ask how the stated spectral range, update interval, stream categories, calibration workflow, and software output map to your analytes, reference procedure, concentration range, and required decision.
The documented configurations provide additional installation questions:
| Configuration | Stated application or condition | Question for your feasibility review |
|---|---|---|
| OrionIR® Lab version | Designed for reactor recirculation loops, pilot-rig side streams, and bench setups | Which arrangement will present representative material during development and independent testing? |
| OrionIR® High-Pressure Version | Our measurement principle is unchanged from the Lab version, allowing evaluation-built chemometric models to transfer to a pressurized installation | What independent evidence will show that the model transfers under your pressure, sampling, and installation conditions? |
Our repeatability test found that OrionIR® held its spectral stability across more than 100 days. That is a stated result from our test, not a performance conclusion for your process or installation. Connect any stability discussion to the measurement interval, environment, maintenance plan, and model lifecycle that your application requires.
We offer ongoing chemometric-model support, including revalidation as feedstock, temperature, or product specifications change. Include that lifecycle question in the review: define which changes trigger reassessment, what samples are needed, and which performance criteria will be applied.
For a Singapore laboratory, pilot rig, or process site serving operations in North America, South America, Europe, or Asia, document the local installation environment and operating requirements rather than assuming that a configuration tested in one setting transfers unchanged to another. Classified-area, continuous-manufacturing, and regulatory questions require their own site-specific review; the documented configuration alone does not resolve them.
Key Highlights
Process FTIR feasibility becomes clearer when you keep the evidence chain in view:
- Evaluate the complete measurement system. Consider the spectrum, calibration, reference method, sampling arrangement, installation, software, and lifecycle support together.
- Define the process domain first. State analytes, concentration extremes, composition changes, temperature, pressure, water, solids, bubbles, and the required decision.
- Anchor calibration to a reliable reference method. Our predicted values cannot be more accurate than the reference method used for training; its reproducibility sets the model’s accuracy floor.
- Test independent samples. Validate the multivariate model against predetermined performance criteria rather than judging it only on calibration data.
- Treat transfer as a technical question. A change in measurement conditions or instrument can require adjustment, additional evidence, or redevelopment.
- Use process samples to resolve uncertainty. We offer to test our technology on your process samples before commitment.
- Plan model support after deployment. We offer ongoing chemometric-model support, including revalidation as feedstock, temperature, or product specifications change.
These points produce a practical decision boundary: proceed to a bounded feasibility review when the process question, representative samples, reference method, installation conditions, and acceptance criteria can be stated clearly. If one of those inputs is still unknown, make it the explicit subject of the next test rather than treating it as a minor detail.
Prepare the evidence package for a feasibility review
A feasibility review works from a defined evidence package rather than a product label. Assemble the information below in one working brief and use it to structure the technical discussion.
1. State the process question. Identify the chemical result you need, the process decision it will support, the required update interval, and the accuracy or error boundary that matters to that decision.
2. Describe the material. List analytes, major components, feedstocks, products, by-products, expected composition changes, concentration range, and the minimum and maximum conditions that must be represented.
3. Document the reference method. Provide the procedure, sampling point, preparation steps, measurement range, repeatability, reproducibility, and the way reference results will be paired with spectra. Our predicted values cannot be more accurate than the reference method used for training.
4. Describe the installation. Record vessel or line location, bypass or recirculation arrangement, pressure, temperature, flow, water, solids, bubbles, phase behavior, optical access, cleaning, maintenance, and data connection. Include how the test material will reach the measurement.
5. Define the model test. Identify calibration samples, independent test samples, concentration coverage, operating-condition coverage, performance measures, predetermined criteria, and the handling of outliers or failed measurements. Independent test data should be compared with criteria established before the result is reviewed.
6. List transfer conditions. Mark every difference between laboratory or pilot evaluation and the intended installation, including instrument, pressure, temperature, humidity, sampling path, composition, and product specification.
We offer to test our technology on your process samples before commitment. That option is most useful when the samples, reference results, process conditions, and decision criteria arrive with enough context to distinguish a model question from a sampling or installation question.
Add ownership to the package as well. Identify who supplies the reference results, who defines the acceptance criteria, who records installation conditions, and who decides whether a changed feedstock or product specification triggers reassessment. A technically clear dataset can still lead to an unclear decision if those responsibilities are left unstated.
Keep the conclusion within the available evidence
General FTIR studies explain how spectra, reference methods, multivariate models, independent samples, and transfer conditions can be evaluated. Our documented OrionIR® capabilities describe stated spectral coverage, measurement frequency, configurations, calibration integration, and process-sample testing. Neither category of evidence establishes suitability for an unspecified process stream.
A stated spectral range does not by itself establish the required analyte accuracy. A high-pressure configuration does not by itself establish acceptance for a particular installation. A calibration model does not by itself establish performance after composition, temperature, pressure, sampling, or instrument changes. Regulatory acceptance and safety classification also require application- and site-specific assessment rather than inference from a general instrument description.
If one of those links remains undefined, identify that link as the next feasibility question instead of converting a general capability into a process result.
The boundary applies equally to positive and negative conclusions. A limited test may show that a particular sample set and arrangement require further work without proving that every related process is unsuitable. Keep the decision tied to the tested analytes, material, conditions, instrument, and criteria.
Define the unresolved questions before requesting a review
A process FTIR review is ready to begin when you can state the process measurement question, the material that represents the process, the reference method and its reproducibility, the concentration extremes, the proposed sampling or optical arrangement, the independent test plan, and the conditions that may require model transfer or redevelopment. That information lets you distinguish a spectral question from a calibration question, a sampling question, or an installation question.
We offer to test our technology on your process samples before commitment, and we offer ongoing chemometric-model support, including revalidation as feedstock, temperature, or product specifications change. A qualified professional must decide individual suitability for a specific process, installation, regulatory setting, or safety requirement. The appropriate next step is a bounded feasibility discussion with the unresolved variables and decision criteria written down in advance.
What should a supplier feasibility discussion establish?
What is process FTIR and how does it differ from laboratory FTIR?
Process FTIR refers to the use of Fourier-transform infrared spectroscopy for real-time chemical measurement in a continuous process environment, capturing spectral data as chemistry evolves. This differs from laboratory FTIR, where samples are collected and analyzed at discrete time points.
What factors should be considered when evaluating the feasibility of process FTIR?
Key factors include the chemical information present in the spectrum, the representativeness of the process material for calibration, the model’s ability to remain useful outside its training data, and the installation’s capability to deliver consistent evidence.
How can I ensure that the calibration model for process FTIR is reliable?
To ensure reliability, the calibration model should be validated with independent test data, and the reference method used for training must be robust and reproducible. Additionally, the calibration set should span the expected concentration range.
What are the implications of using a reference method in process FTIR?
The reference method sets a boundary for the accuracy of the spectroscopic model; predicted values cannot exceed the accuracy of the reference method used for training. It’s crucial to document the reference procedure and its conditions.
What steps should I take to prepare for a feasibility review of process FTIR?
Prepare a feasibility review by stating the process question, describing the material and its characteristics, documenting the reference method, detailing the installation setup, defining the model test plan, and listing any transfer conditions that may affect the model.
What should I do if my process conditions change after calibration?
If process conditions change, such as temperature, pressure, or composition, you may need to reassess the calibration model. This could involve recalibration or validation with new independent samples to ensure the model remains valid.
How does the installation setup affect the performance of process FTIR?
The installation setup, including the sampling arrangement and optical access, directly impacts which material reaches the measurement. It’s essential to ensure that the setup allows for representative sampling under the expected operating conditions.