Analytic Journal is growing – your feedback shapes the platform.Send feedback
Category: PAT | Measurement Technology | PAT Fundamentals
Author: Dusko Kadijevic, Publisher Analytic Journal
Last updated: May 2026
Reading time: approx. 8 minutes
Target audience: Process engineers, instrumentation engineers, PAT specialists, plant designers
Inline, Online and At-line describe the spatial and process-related relationship between the measurement point and the running process. They define where and how a measurement is carried out — and directly determine response time, sample representativeness, personnel effort, maintenance requirements and suitability for different process environments.
The term In-situ is widely used in practice as a more direct form of inline measurement and is not always listed as a separate main category. Where the distinction matters, it is explained in the Inline section.
The choice of the right measurement strategy is one of the most fundamental decisions in PAT — and one of the most common sources of error in practice.
Pure quality control does not create value — the industry knows this. The real benefit of the right measurement strategy lies elsewhere: faster processes, lower energy consumption, reduced raw material use, less waste and rework. Measuring at the right point with the right method enables active process control rather than retrospective inspection.
That is the business case behind inline and online — not compliance, but competitiveness.
The economic potential of PAT — process optimisation, resource efficiency and automation — is covered in depth in the article "What is PAT? — PAT as the Key to Modern Production."
| Characteristic | Inline (incl. In-situ) | Online | At-line |
|---|---|---|---|
| Measurement location | Directly in the process stream, no sample path | Extractive sampling, continuous | Sample taken near the process, discontinuous |
| Sample conditioning | Not required | Required (filter, cooler, pump) | Manual or automated |
| Response time | Seconds | Minutes | Minutes to hours |
| Sample representativeness | Very high | High, when sampling is correctly designed | Depends on sampling quality |
| Personnel effort | Low | Medium, mainly for conditioning maintenance | High — sampling and measurement often manual |
| Maintenance effort | Medium to high | High, mainly in the sampling system | Lower |
| Capital cost | Medium to high | Medium | Lower |
| Aggressive media | Conditional | Yes, with appropriate conditioning | Yes |
| Typical devices | TDLS, pH probe, NIR probe, ZrO₂ | NDIR, GC, paramagnetic, TOC | Portable instrument, laboratory analyser |
An inline sensor measures directly at the process — without external sample extraction and without downstream sample conditioning. The sensor is physically located in the medium or measures immediately at the process point.
Typical inline measurement principles include TDLS, pH electrodes in reactors, NIR probes in granulate or liquid streams, and conductivity sensors in purified water applications.
Response time is fastest because the sensor reacts directly to process changes. Since no sample conditioning is required, there are fewer failure sources and lower maintenance demand. The delay caused by sample transport is eliminated entirely. Sample representativeness is very high because exactly what is present in the process is measured.
The most common cause of failure in inline measurements is that the process medium was not sufficiently investigated before sensor selection.
In practice this means: a sensor is selected for inline operation because it appears technically suitable — but the actual process conditions have not been fully analysed. Typical problems include:
Consequence: Before any inline decision, a complete analysis of process conditions is required — not just temperature and pressure, but also particulates, reactive components and material compatibility across all process states, including start-up, shutdown and CIP (Cleaning-in-Place — automated cleaning of the plant without dismantling, standard practice in pharmaceutical and food manufacturing).
Note on In-situ: In technical literature and practice, in-situ is frequently used as a synonym for inline. The relevant distinction arises when a measurement principle operates without direct media contact — for example TDLS, where the laser beam traverses the gas volume without extracting a sample. This is technically advantageous for aggressive, hygiene-critical or reactive media. In most other contexts, inline and in-situ are effectively equivalent.
An online system continuously extracts a representative sample from the process, conditions it and feeds it to the analyser. Conditioning typically includes filtration, cooling, pressure reduction and drying — depending on the medium and analyser.
The analyser itself is not located in the process but in an analyser shelter or cabinet, protected from harsh process conditions.
Online enables the use of sensitive and complex analysers that are not designed for direct process conditions. Maintenance work on the analyser can often be carried out without interrupting the process. The approach is also suitable for aggressive, corrosive or heavily loaded media, provided the sample conditioning and material selection are designed accordingly.
The sample conditioning system is the most critical element. The reliability of the overall system is only as good as the sampling system. A blocked filter, a failed pump or a leaking sample line can render the analyser blind — without the analyser itself indicating any fault. This is an undetected failure in measurement technology terms.
A further disadvantage is time delay: between the process event and the measured value lies the transport time of the sample through the line. With long sample paths this delay can amount to several minutes — critical for fast control tasks.
The medium may also change along the way from the process to the analyser — through condensation, chemical reaction with line materials or degassing. Particularly critical for reactive gases or media containing volatile components.
| Component | Function | Typical Failure Mode |
|---|---|---|
| Pre-filter / coarse filter | Particulate removal, particle protection | Blockage from high dust loading |
| Fine filter | Protection of the analyser | Blockage, bypass leakage |
| Sample gas cooler (Peltier cooler, recirculating cooler) | Moisture removal from sample gas, dewpoint reduction | Condensate accumulation, cooler failure |
| Condensate drain (condensate pump, siphon) | Removal of condensate from the cooler | Blockage, pump failure |
| Sample gas pump (diaphragm pump, ejector pump) | Gas transport from process to analyser | Diaphragm wear, pressure loss |
| Liquid sample pump (peristaltic, piston pump) | Transport of liquid samples in extractive liquid analytics | Tubing wear, leakage |
| Heated sample line | Prevents condensation along the sample path | Heating failure, leaks |
| Flow meter / controller | Constant sample gas volume flow to the analyser | Drift, blockage, failure |
| Pressure regulator | Adjustment of sample pressure to analyser inlet | Diaphragm wear |
Important note for SIL applications: In safety-relevant online systems, the entire sample conditioning system — filter, pump, flow meter and line — must be included in the SIL assessment. A filter failure that renders the analyser blind is a safety-relevant failure, even if the analyser itself reports no fault.
At-line describes a measurement that takes place spatially close to the process but outside the actual process stream. A sample is taken manually or automatically, brought to the measuring device and analysed there. The measuring device is not integrated into the process.
Examples include a portable NIR instrument at the production line, a laboratory analyser directly beside the plant, or an automated sampling station with discontinuous analysis.
At-line has the lowest capital cost among the three strategies — no complex inline installation and no elaborate sampling system are required. The approach is flexible because the same instrument can be used for multiple measurement points. Maintenance and calibration are simpler since the device is not integrated into the process.
The greatest disadvantage is time delay. Between sample collection and measurement result, minutes to hours pass depending on the process and analysis requirements. For control tasks, at-line is therefore generally not suitable.
A frequently underestimated cost factor is personnel effort. At-line is not continuously automated: someone must collect the sample, bring it to the measuring device, operate the instrument, read the result and document it. In three-shift operation this means qualified personnel must be available around the clock. The lower capital cost compared to inline or online is often offset by ongoing personnel costs.
With automated at-line systems, personnel effort decreases but does not disappear entirely: calibration, maintenance of the sampling system and result validation remain manual tasks.
Sample representativeness also depends heavily on the sampling process. If the sample is not correctly collected, transported and prepared, the measurement result is unreliable — regardless of how good the measuring device itself is. Particularly with inhomogeneous media such as suspensions or multiphase systems, sampling is a discipline in its own right.
At-line also does not deliver continuous measured values. Relevant process changes between two measurements can go unnoticed.
Off-line refers to classical laboratory analysis: a sample is taken from the process, transported to a laboratory and analysed there under controlled conditions. The laboratory is spatially separate from the process — the time delay between sampling and result ranges from hours to days.
Off-line is not a PAT method in the strict sense, but it is the foundation on which all other measurement strategies are built. Calibration models for NIR probes are developed off-line. In the United States, product release decisions under FDA 21 CFR Part 211 are based on validated laboratory methods. In the UK, MHRA Good Manufacturing Practice requirements follow the same principle. Validation studies for inline and online systems require off-line reference measurements.
A common misconception in PAT projects is that off-line should be replaced by inline or online. That is the wrong approach.
The right question is not "How do I replace my off-line measurement?" — but: "Where do I measure what, to achieve optimal process control and automation?"
Off-line remains indispensable for tasks that require it:
What inline and online can deliver — and off-line cannot — is continuous real-time process control. Not because off-line measures less accurately, but because the time delay is too great for control tasks.
| Task | Appropriate Strategy |
|---|---|
| Real-time process control | Inline / Online |
| Early detection of process deviations | Inline / Online |
| Quality assurance and product release | Off-line (+ At-line as complement) |
| Calibration of process sensors | Off-line as reference |
| Validation studies | Off-line |
| Raw material inspection on receipt | At-line / Off-line |
| Finished product testing | Off-line |
Legend: ✅ Well suited | ⚠️ Conditionally suited | ❌ Not suited
| Requirement | Inline | Online | At-line | Off-line |
|---|---|---|---|---|
| Fast process control (< 1 min) | ✅ | ⚠️ | ❌ | ❌ |
| Continuous monitoring | ✅ | ✅ | ❌ | ❌ |
| Aggressive / corrosive media | ⚠️ | ✅ | ✅ | ✅ |
| High particulate loading | ⚠️ | ✅ | ✅ | ✅ |
| High-sensitivity analytics | ❌ | ✅ | ✅ | ✅ |
| Low personnel effort | ✅ | ✅ | ❌ | ❌ |
| GMP-compliant product release | ❌ | ❌ | ⚠️ | ✅ |
| Calibration / validation | ❌ | ❌ | ⚠️ | ✅ |
| SIL-relevant safety monitoring | ✅ | ⚠️ | ❌ | ❌ |
| Multicomponent measurement | ⚠️ | ✅ | ✅ | ✅ |
| Low capital cost | ❌ | ⚠️ | ✅ | ⚠️ |
| High sample representativeness | ✅ | ✅ | ⚠️ | ⚠️ |
| Flexible use across multiple points | ❌ | ❌ | ✅ | ✅ |
Pharmaceutical industry — granule moisture measurement: NIR probe inline in the granulator — direct real-time measurement of residual moisture without sample extraction. Suitable when the granulate contains no reactive components that could coat the probe. At-line is an alternative when inline conditions are too aggressive.
Chemical industry — synthesis gas monitoring: Online GC for multicomponent measurement — synthesis gas contains H₂, CO, CO₂ and CH₄ in varying ratios. TDLS inline can be added for fast O₂ safety monitoring in parallel.
Water treatment — TOC measurement: Online TOC analyser with extractive sample conditioning — standard in purified water monitoring, including applications governed by USP <643> in the US and EP 2.2.44 in Europe. Inline is not practical here as TOC measurement requires conversion of the carbon content.
Food and beverage — Brix measurement: Inline refractometer directly in the process pipe — fast, hygienic and without sample extraction. CIP capability of the sensor is a mandatory requirement.
Refineries — H₂S monitoring: Online with electrochemical sensor and cooled sample line — H₂S is too corrosive and toxic for most direct inline applications. OSHA permissible exposure limits and EPA risk management requirements make continuous monitoring mandatory at relevant sites.
| Standard / Source | Content |
|---|---|
| FDA PAT Guidance (2004) | Foundational definition of inline, online, at-line and off-line in the pharmaceutical context — the reference document for PAT terminology |
| ASTM E2363 | Standard terminology for process analytical chemistry — definitions of measurement positions |
| USP <1058> | Analytical instrument qualification — relevant for at-line and off-line instruments in US pharmaceutical manufacturing |
| ISO 15839 | Field devices for water analysis — terminology definitions for measurement positions |
| NAMUR NE 107 | Self-monitoring and diagnostics of field devices — relevant for inline and online systems |
| IEC 61511 | Safety instrumented systems for the process industry — measurement positions in safety loops |
| 21 CFR Part 211 | US FDA current Good Manufacturing Practice — governs laboratory controls and product release testing |
| MHRA GMP Guidelines | UK equivalent to FDA GMP — governs pharmaceutical manufacturing and quality control in Great Britain |
What do inline, online and at-line mean?
Inline means: the sensor measures directly at the process without sample extraction. Online means: a sample is continuously extracted, conditioned and fed to an analyser. At-line means: a sample is taken discontinuously and analysed close to the process. The choice between the three strategies determines response time, maintenance effort and suitability for different process conditions.
When is inline better than online?
Inline is better when fast response times are required for process control, when the medium does not cause deposits on the sensor, and when material compatibility between sensor and medium has been fully verified. Online is better when the medium is too aggressive for direct sensor contact or when sensitive analysers are needed that cannot withstand direct process conditions.
What is the most common error in inline measurements?
The process medium is not fully analysed before sensor selection. In practice this means: particulates, reactive components or deposits in the process gas or liquid are not accounted for. In TDLS systems, solids can interfere with optical windows and cause signal loss. In liquid probes, deposits can bias measurements without triggering a fault alarm — a particularly dangerous undetected failure.
Why is sample conditioning so critical in online systems?
The reliability of an online system is only as good as its sampling system. A blocked filter, a failed pump or a leaking line renders the analyser blind without the analyser itself reporting a fault. In safety-relevant applications, the entire sample conditioning system must therefore be included in the SIL assessment — not just the analyser.
Is at-line suitable for process control?
Generally not. At-line delivers discontinuous measured values with a time delay of minutes to hours — too slow for control tasks. At-line is better suited for quality control, release decisions and process monitoring where no real-time signal is required.
What does off-line mean and how does it differ from at-line?
Off-line means: the sample is taken and analysed in a separate laboratory — spatially and temporally distant from the process, with a delay of hours to days. At-line is spatially close to the process and considerably faster. Off-line is the fourth category in PAT terminology alongside inline, online and at-line.
Should off-line be replaced by inline or online?
No — this is a common misconception in PAT projects. Off-line often measures more accurately using certified methods. What inline and online provide is continuous real-time process control — something off-line cannot deliver due to its time delay. The right question is not "How do I replace off-line?" but "Where do I measure what, to achieve optimal process control?" Each strategy has its role: inline and online for control and monitoring, off-line for release, calibration and validation.
Which measurement strategy requires the most personnel effort?
At-line — clearly. Since at-line is not continuously automated, qualified personnel must collect the sample, bring it to the instrument, operate the device and document the result. In three-shift operation with frequent measurement intervals, this adds up to a significant cost factor. Online and inline run largely automatically; personnel effort is limited to maintenance and calibration.
How do inline and in-situ differ?
Both strategies measure without sample extraction directly at the process, which is why the terms are often used interchangeably. The more precise distinction is: inline means the sensor works in direct contact with the medium. In-situ means the measurement occurs without direct media contact of the sensor — for example TDLS, where the laser beam traverses the gas volume without extracting a sample. This distinction is relevant for questions of material compatibility, contamination risk and CIP suitability.
Which measurement strategy is suitable for SIL-relevant applications?
Inline is usually preferred because fewer components are present in the safety loop and therefore fewer failure sources exist. In online systems, the entire sample conditioning system — filter, pump, flow meter and line — must be included in the SIL assessment. At-line is not suitable for safety-relevant real-time monitoring because measurement is not continuous.
The choice between inline, online and at-line is not a technical detail — it is one of the most fundamental strategic decisions in PAT. It determines response time, sample representativeness, maintenance effort, capital cost and suitability for safety-relevant applications.
The most important lesson from practice: an inline measurement is not automatically better simply because it is more direct. When the medium causes deposits, material compatibility has not been fully verified, or process conditions are too aggressive, online with a well-designed sample conditioning system is often the more reliable choice.
And for online systems the reverse applies: the analyser is only the final stage. The sample conditioning system is the heart of the installation — and the most common place where reliability is lost.
Analytic Journal — The Portal for PAT
© 2026 Analytic Journal / Dusko Kadijevic. All rights reserved.
This article is protected by copyright. Short excerpts with source attribution and an active link to analyticjournal.com are permitted. Full or substantial reproduction — including in digital form, by AI systems or automated processes — requires the prior written consent of the publisher. Enquiries: info@analyticjournal.de
Related topics on Analytic Journal:
Editorial Note
This article has been written to the best of our knowledge based on publicly available technical information, standards, and manufacturer data. Specifications may change — for binding information, consult current manufacturer datasheets and relevant standards bodies. Analytic Journal produces editorial content independently of manufacturers and advertising partners.
Author: Dusko Kadijevic — 20 years of experience as a PAT practitioner in the process industry
Dusko Kadijevic | Publisher Analytic Journal | analyticjournal.com
Written May 2026. All content formulated in our own words, created with AI assistance.
First published on 8 June 2026 on analyticjournal.com