Analytic Journal is growing – your feedback shapes the platform.Send feedback
10 June 2026
Category: Process Analytical Technology | Instrumentation Strategy | Case Studies
Author: Dusko Kadijevic, Publisher Analytic Journal
Last updated: June 2026
Reading time: approx. 14–16 minutes
Target audience: Process engineers, PAT specialists, production managers, chemical engineers, quality assurance professionals
Root-cause-oriented PAT specification is the systematic identification of the process variable or process condition that causes a problem — rather than monitoring the visible symptoms of that problem. The goal is to control the cause, not to detect the symptom.
Process Analytical Technology (PAT) is often equated in industry with introducing faster online measurements. This view falls short.
A measurement only becomes valuable when it captures a variable that is truly decisive for understanding, controlling, or preventing a critical process state.
This is exactly where a common source of misinvestment lies: teams specify measurement systems based on the symptom, not based on the cause.
This leads to technically clean but process-wise weak solutions.
Whoever recognizes a problem only at the end of the process chain can usually only manage it — not prevent it.
A typical pattern in the process industry is concentration on what is immediately visible.
When a facility causes H2S odor, the idea of an H2S measurement at the reactor outlet seems obvious. This idea is understandable:
The issue with this logic is not its technical feasibility, but its position in the process.
A measurement at the reactor outlet only recognizes that the reaction has already taken place.
| Feature | H2S Detection at Outlet | pH Control in Reactor |
|---|---|---|
| Detection timing | After H2S formation | Before H2S formation |
| Response type | Reactive (responds to symptom) | Preventive (avoids cause) |
| Information content | Binary or event-based | Continuous and quantitative |
| Process stability | Limited | Active control possible |
It is therefore reactive, not preventive.
For PAT, this is only sufficient if the actual cause of the problem cannot be controlled.
A chemical manufacturing company faced a recurring problem: H2S gas was escaping from a reactor.
The consequences were immediate:
The facility already possessed an exhaust gas scrubber system — leftover from an earlier process — that could theoretically neutralize H2S.
The first technical response was intuitive: Online H2S measurement at the reactor outlet + automatic scrubber activation.
Logical. Understandable. And it was wrong.
A deeper analysis of the process led to a different question:
Not the H2S value itself, but the pH inside the reactor was the actual driver of gas formation.
In other words:
This changed the entire specification logic:
Old question: "How do we detect H2S faster?"
New question: "Which process variable do we need to control so that H2S never forms critically in the first place?"
At this point, a second, often underestimated problem emerged.
The facility had already attempted to measure pH online under harsh process conditions.
The previous solution required:
This was labor-intensive, failure-prone, and only practical to a limited extent during ongoing operations.
From this experience, the conclusion was drawn: "pH measurement is fundamentally not feasible in this process."
This conclusion, however, was time-bound.
The state of the art had evolved:
Retractable probe systems with automatic return to a calibration chamber now allow continuous pH measurement with automatic calibration cycles — without opening the reactor, without regular manual intervention.
The decisive insight:
An old technology judgment must not be transferred uncritically into the present.
If you assessed a technology as "impractical" decades ago, you should reassess it today.
In this case, pH measurement was fundamentally possible — but only the use of a resistant process electrode type made it practical under the harsh conditions. Manually operated electrodes were not cleaned and calibrated regularly, due to the effort involved, which drastically shortened their service life through inaccuracy. The extreme process conditions — H2S itself also damages the electrode — occurred too frequently, further degrading service life. A vicious cycle: the less maintenance was performed, the faster the probe failed, and the more the impression grew that pH measurement was fundamentally impractical here.
If pH changes trigger H2S formation, then pH is the actual control variable.
Comparison of the two approaches:
| Approach | What Is Measured | Where | What Is Controlled |
|---|---|---|---|
| H2S Detection | H2S concentration | Reactor outlet (downstream) | The symptom |
| pH Control | pH value | Inside reactor (in-process) | The cause |
H2S measurement at the end of the process answers only one question:
"Has the problem already become visible?"
pH measurement inside the reactor answers the more important question:
"Are process conditions already drifting toward malformation?"
This is not a technical but a strategic problem of specification.
The measurement variable is not the most directly visible, but the most causally effective.
The solution consisted of automated pH measurement with a retractable probe and cyclical self-calibration:
For practice, three levels are decisive:
Only when these three levels fit together does a measurement become a reliable control instrument.
This pattern is not limited to H2S and pH.
| Industry | Symptom (visible) | Cause (hidden) | Right Measurement Variable |
|---|---|---|---|
| Pharma | Late product deviation | Mixing sequence, pH, or temperature profile | Control upstream variable |
| Chemicals | Byproduct formation | Inadequate temperature or dosing control | Process parameter, not product |
| Water | Discharge limit exceedance | Inadequate dosing or inlet quality | Measure inlet and dosing |
| Food | Taste/texture deviation | Raw material variability or process drift | Raw material + process condition |
The common denominator is always the same:
First determine the cause, then select the appropriate measurement variable.
Whoever reverses this sequence often ends up with a measurement that is technically interesting but too late from a process perspective.
Three fundamental questions can be derived from this case, to be answered in this order:
Not every observable variable is also a meaningful control variable.
The decisive question is: Which variable, if kept stable, prevents the problem?
This requires deep process understanding, conversations with plant operators and chemists.
An ideal measurement variable is of little use if it cannot be reliably captured in the process.
Therefore, measurability must be assessed under actual conditions — not just in the laboratory or on the manufacturer's datasheet.
Only now does it come down to:
These questions are important, but they come third, not first.
What is root-cause-oriented PAT specification?
Root-cause-oriented specification is the systematic identification of the process variable whose control prevents a problem — not the monitoring of visible symptoms. The goal is prevention, not detection.
What is the difference between symptom and cause in this context?
A symptom is what you observe (odor, off-spec product, deviation). A cause is the process condition that creates the symptom. If controlling a variable prevents the symptom, you have identified the cause.
Why do most PAT projects focus on symptoms instead of causes?
Symptoms are visible, urgent, and easy to measure. Causes are hidden, technical, and require investigation. Teams naturally gravitate toward the visible problem. Additionally, symptom solutions often have obvious measurement approaches, while cause solutions require conversational processes.
In which situation is H2S measurement still the right choice?
H2S measurement is correct when the upstream cause cannot be controlled, or when multiple uncontrollable factors lead to H2S formation. In that case, early detection and exhaust gas treatment is the only practical path. That was not the case here.
What are retractable probe systems and how do they work?
Retractable probes (also: Automatic Probe Withdrawal, APW) are measurement probes that automatically withdraw from the process into a calibration chamber. They self-calibrate against known standards and then return to the process. This allows continuous measurement without manual intervention or reactor interruption.
How long should you wait with pH measurement under harsh conditions before concluding it doesn't work?
Modern retractable systems require a validation phase of 4–8 weeks of parallel laboratory analysis. If the online measurement consistently aligns with laboratory results, it is ready for control. If it doesn't track in parallel, the cause must be investigated (fouling, calibration issues, wrong measurement variable).
What is the difference between monitoring and control?
Monitoring: You observe a value and react to it (alarm = activate scrubber).
Control: The system automatically adjusts a process input to keep the value within range (pH drops = system automatically doses base).
Control is proactive. Monitoring is reactive.
Can old technology judgments still be valid in current practice?
Yes, but not automatically. If you assessed pH measurement as "impractical" decades ago, you should ask the question again today. Retractable probes, automatic calibration, and cloud integration have evolved significantly. An old judgment must be re-examined.
What effort is needed to switch from symptom measurement to cause measurement?
It's less a question of time and more a question of who's in the room. If you have the right people at the table — someone who understands the chemistry and someone who understands the physical reality inside the reactor (the process engineering perspective) — root cause analysis can happen quickly, often within days. What typically delays projects isn't the analysis itself, but the absence of one of these two perspectives. A purely instrumentation-focused team without deep process or chemistry expertise will struggle to identify the cause, no matter how much time is given.
What is the most common mistake in PAT specifications?
The measurement variable is specified before the process cause is understood — often because the project team lacks either the chemistry expertise or the process engineering perspective needed to identify the real driver. Bringing both perspectives together early is more valuable than any amount of additional project time. Without them, teams default to measuring the visible symptom because that's the only thing they can confidently define.
Spectroscopic Measurement Principles (NIR, Raman, IR):
This article focuses on the strategy of measurement variable selection, not on physical measurement principles. A deeper treatment of spectroscopy and its limitations under process conditions can be found in the upcoming article "From NIR to HPLC: Which Measurement Principle for Which Process?"
Validation and Regulatory Compliance (21 CFR Part 11, ATEX):
While validation is critical, this article's focus is on the strategic selection of the measurement variable, not regulatory implementation. Compliance details are covered in "PAT Validation and Regulatory Requirements."
Data Management and Real-Time Analysis:
Once reliable measurements exist, the data must be collected, analyzed, and implemented into control loops. This is the subject of a separate article: "Data Architecture for Online Process Control."
PAT projects do not fail due to insufficient measurement technology, but due to inadequate problem definition.
Whoever begins with the visible symptom often builds a solution system that only addresses the consequence.
Whoever begins with the cause can develop a measurement and control strategy that actually improves the process.
The added value lies not only in avoiding the problem, but in a better understanding of production reality.
Three central lessons from this case:
PAT must address the cause, not the symptom. H2S at the reactor outlet is an important warning signal. It is no substitute for controlling the actual process variable. If pH drives H2S formation, then pH is the correct control variable — not H2S.
Technical possibilities continue to evolve. What was not practical in a process decades ago may very well be achievable today with appropriate measurement technology. Good PAT specification therefore requires not only process understanding, but also regular updating of one's own technological knowledge base.
The right sequence determines success. Question 1: Which variable do I need to control? Question 2: Can I measure it reliably? Question 3: Which technology makes that practical? Whoever reverses this sequence buys the measurement first and looks for a problem it solves afterward.
Primary Sources (Case Study & Process Data)
Technical References
Industry Standards and Best Practices
Dusko Kadijevic is a PAT specialist with 15+ years of experience in process analytical technology across the pharmaceutical, chemical, and specialty materials sectors. He is the founder of Consulting DK and Publisher of Analytic Journal. His work focuses on bridging the gap between analytical science and process control.
© 2026 Analytic Journal / Dusko Kadijevic. All rights reserved.
Short excerpts with source attribution and an active link to analyticjournal.de 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
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