pH: REDUCING ERROR COMPARING IN-LINE VERSUS

OFF-LINE TECHNIQUES

David Bigalke
National SalesM anager
M ettler-ToledoIngold,Inc.
36M iddlesexTurnpike
Bedford,M A 01730

KEYW ORDS

pH Electrode,In-line,Off-line,M aintenance,Calibration,Sampling

ABSTRACT

In-line processpH measurementsare oftencompared tooff-line grabsamples.The off-line technique is

commonlytreated asthe “gold standard”while the in-line measurementequipmentisusuallyheld up to
a highdegree ofscrutiny.The in-line reading isalwayssuspectwhent he t
woresultsdonotagree.

The purpose oft hispaper ist

oexplaint
he cause ofdiscrepanciesbetweenthe twotechniquesand to
focusonmeansfor minimizing errorsbetweenin-line and off-line pH measurements.Among topicsto
be discussed include:

x Sample pointlocation
x The importance oftemperature
x Timing
x M easurementtechnique
x Electrode maintenance
x Ground loops
x Understanding the limitations

Having a thoroughunderstanding ofthe measurementprocesses,and paying attentiontodetailsof

calibration,sampling,and measurement,will minimize off-line analysisrequirementsand deliver anin-
line pH measurementyoucanrelyon.

Copyright2005byISA -The Instrument ation,Systemsand AutomationSociety.

Presented at50thISA AnalysisDivisionSymposium; http://www.isa.
org
 INTRODUCTION

Accurate measurement of pH is critical to optimize product yield, maintain product quality, and provide
personal and environmental protection. In-line pH measurement provides immediate feedback of
process conditions allowing for tight process control. In-line measurements are often periodically
checked by use of off-line measurements. It is important to understand and quickly resolve the cause of
any observed discrepancy between in-line and off-line readings. The main focus of the following primer
is to outline how electrode technology and sampling technique can work together and how to best
minimize errors.

There are many manufacturers of good quality off-line pH measurement equipment. The first
assumption of this paper is that the user is equipped with a good bench top or portable pH system and
pH electrodes. It is important to note that portable pH meters are available for general purpose or
hazardous area use (FM Class I, Div. 1). Be sure to use correctly rated equipment for the area of use.
The second assumption is that the electrode has been properly maintained. The guidelines below serve
to direct the user to the best use of this equipment for proper evaluation of the integrity of an in-line
measurement.

PROPER MAINTENANCE TECHNIQUE

The off-line pH meter and the electrode should be held to the same standards as the in-line measurement
equipment. First, the off-line meter should be checked with a good quality pH simulator (Refer to
Figure 1). The purpose of the simulator is to not only verify the operation of the pH meter (ability to
read a millivolt signal)but as a test of the electrical connections and cables. The only way to truly test
the electrical connections and cables is with a simulator that provides both a high and a low impedance
millivolt output. High impedance simulators provide the same type of signal you would expect from a
pH electrode. A difference in readings between the high and low impedance millivolt output can
indicate a problem with the cable or one of the connections. This simulator test should be done annually
at a minimum.

FIGURE 1–pH SIMULATOR

Maintenance such as storage, cleaning, etc. is defined by the manufacturer of the electrode and meter.
The “health check” characteristics of the glass Ag/AgCl pH electrode, however, are universal. There are
three key parameters that should be examined: the offset, the slope and the response time. The
following is an individual discussion of each parameter.

Copyright 2005 by ISA - The Instrumentation, Systems and Automation Society.

Presented at 50th ISA Analysis Division Symposium; http://www.isa.org
ELECTRODE OFFSET

The millivolt reading of an electrode in a pH 7 buffer is called the electrode offset. Ideally this reading
is zero. Some electrodes come with a quality certificate indicating the electrode is +/- 0.25 pH units or
15 mV as new. Rarely does this number exceed 5 millivolts. This reading will shift once the electrode
is contaminated with process, whether it is an in-line or off-line electrode. The degree of shift the user
can tolerate is dependent on both the tolerance for error and the instrument. Some instruments will not
allow the calibration of an electrode with high offset values, e.g. exceeding 100 millivolts. An offset of
this degree is indicative of something wrong with the electrode, such as a fouled j unction or poisoned
reference. It is a warning sign that the electrode is unstable and should not be used until it is restored via
cleaning, reference solution change, etc. It is important to set an offset limit for your plant.

SLOPE

A reasonable “out of the box” standard for electrodes is a slope of >98%. Slopes >101% usually
indicate something is wrong in the calibration e.g. technique or buffers. Electrode failure is usually
accompanied by a decreasing slope, never an increasing slope. The slope of the electrode will decrease
as the pH glass ages. Most instruments have a built in or user defined limit for slope. 85% would be a
good starting point for a lower limit, although each plant should develop their own rules based on their
process. The plant may also want to define cleaning stages e.g. 90%.

RESPONSE TIME

The response time of the electrode can decrease as the active layer on the pH glass ages or becomes
chemically bound or physically coated. Many off-line instruments contain a feature that automatically
locks in the reading based on a rate of change per unit time. This is often desired by users because it
takes the manual guess work out of when to take the pH reading. In this case, the response time of the
electrode can have an impact on when the instrument locks in. Response time standards should also be
set based on the process and the sample pH change with time for this reason. Many samples will
dissolve or release soluble gases with time thereby requiring a very fast electrode response time. A
standard out of the box specification for a process electrode is 20 seconds (98% response between pH 4
and pH 7).

BUFFER SELECTION

Buffer selection is a topic that includes: choosing a good manufacturer, choosing the correct buffers and
maintaining good laboratory practice when using the buffers. First and foremost, it is important to select
a buffer manufacturer that produces buffers to some universal standard such as NIST. Ideally you will
select buffers that bracket the range you are measuring. Sometimes this is not possible, if the process
has a wide enough range. Most important is to realize that the longer and more frequently you use a
buffer, the more accuracy you will sacrifice due to contamination and degradation. Fresh certified
buffers play an important role in system accuracy.

Copyright 2005 by ISA - The Instrumentation, Systems and Automation Society.

Presented at 50th ISA Analysis Division Symposium; http://www.isa.org
 UNDERSTAND THE LIMITATIONS

Perhaps the most important point in this primer is to understand the limitations of the pH measurement
technology. A good rule of thumb is that you should be able to maintain 0.2 pH units of agreement
between your off-line and in-line measurement in most applications [1]. Not 0.02 pH, but 0.2 pH. 0.1
pH is obtainable and in extreme cases, even 0.05 pH can be achieved with use of great care (usually
through the use of liquid reference flowing-junction electrodes). This is very difficult to accept if you
have a process specification calling out +/- 0.05 pH units.

Errors associated with pH measurement are additive. Some errors are obvious: the field transmitter
usually has a 0.02 pH unit error. Each buffer has a 0.01-0.02 pH unit error. Assuming perfect
calibration technique, buffers 100% to specifications, no offset introduced by the interaction of the
electrode and the process solution, no grounding errors and no error in sampling, results in 0.04-06 pH
units predicted accuracy in an ideal world. Most industrial users realize that their plants are far from
ideal and some of the errors discussed above will be experienced.

An experiment is recommended to demonstrate the above point. Calibrate the laboratory and/or the
industrial pH system with two buffers e.g. pH 4 and pH 7. After calibration, measure a buffer that lies
between them e.g. pH 5. Examine the error in the measurement of the check buffer alone. This is the
error in an ideal environment. The error observed will, in most cases, be greater than the 0.01 pH units
that many bench top units boast.

The interaction between the process solution and the electrode is the most significant source of error
between in-line and off-line pH measurements.

The most common pH sensor today is the combination silver/silver chloride electrode [1]. The
combination electrode is really two half cells, the pH measuring element and a reference element (Figure
2) combined into a single body.

pH REFERENCE COMBINATION
MEASURING ELECTRODE ELECTRODE
ELECTRODE

FIGURE 2– COMBINATION ELECTRODE

pH is a differential measurement between these two sensors. The differential is created by the
electrochemical reaction between the pH sensing element and the reference element. The pH sensing
element detects the activity of H+ ions through the transfer of electrons across the specially formulated
Copyright 2005 by ISA - The Instrumentation, Systems and Automation Society.
Presented at 50th ISA Analysis Division Symposium; http://www.isa.org
pH sensitive glass, and its voltage changes in proportion to pH. The function of the reference electrode
is to provide a stable voltage to compare against the sensing element. The reference element must be
electrically connected to the process liquid in order to complete the electrical circuit. This connection is
most commonly made via a porous junction.

Most laboratory electrodes and some process electrodes utilize a reference electrolyte solution that
physically flows across the junction into the process at a very slow rate. Most in-line electrodes rely on
only an ionic connection with the process and no flow of bulk electrolyte across the junction.

Ionic strength of the process is different than that of the reference electrolyte in nearly every case. This
difference creates an ionic gradient between the reference solution and the process solution. This ionic
gradient adds (or subtracts) millivolts to the pH measurement that are not accounted for in the
calibration. This voltage is called the junction potential (E2, Figure 3) [2]. The value of the junction
potential is calibrated in buffer but will change in the process.

E MEASURED W here:
E1 = the potential developed at the reference
pH METER
electrode silver/silver chloride wire.
E2 = the “junction potential”
REFERENCE E3 = the potential developed at the “pH cell”
ELECTROLYTE E1 silver/silver chloride wire.
E4 = the pH potential developed at the inner
E2 surface of the pH glass bulb
REFERENCE
JUNCTION
E5 = the pH potential developed at the outer
E3 surface of the pH glass bulb
INNER BUFFER
E4 E Measured = sum of E1 through E5
E5

FIGURE 3– MEASURED VOLTAGE IS THE SUM OF SEVERAL INDIVIDUAL VOLTAGES

An example of junction potential for a given electrode is as follows for hydrochloric acid (HCl) at 25 oC
(Table1) [3].

TABLE 1 – JUNCTION POTENTIAL AS A FUNCTION OF

HYDROCHLORIC ACID CONCENTRATION

HCl Molarity JunctionPotential

1M 14.1 mV
0.1 M 4.6 mV
0.01 M 3.0 mV

The junction potential that is developed is not only dependent on the process composition it is also
dependent on the construction of the reference cell and the condition of the reference junction. Ceramic

Copyright 2005 by ISA - The Instrumentation, Systems and Automation Society.

Presented at 50th ISA Analysis Division Symposium; http://www.isa.org
diaphragms versus Teflon may exhibit different ion mobility for example. A liquid filled versus a gel
filled electrode would be another example. The design of the laboratory electrode is often different from
the process electrode. The effect of these differences has to be taken into consideration rather than
accepting the laboratory electrode as the “gold standard”.

One of the easiest ways to decrease error between in-line and off-line measurements is to use the same
electrode make/model in both the process and the lab. Most process electrodes can be easily interfaced
with popular bench top instruments. The best solution is to set up an off-line measuring station with the
same type of equipment used in-line.

If the above is not an acceptable solution, then the error needs to be determined up front via
experimentation and built into the tolerance between in-line and off-line measurement.

SAMPLE POINT LOCATION

It is very important to account for any differences between the in-line measurement and your grab
sample location. Solutions that are viscous, detergents for example, will have a tough time mixing,
especially where there is laminar flow. A downstream sample point versus an upstream in-line
measurement could see significantly different results. Another example would be chemistry that takes
time to react. Slaked lime added to a process stream can vary reaction time based on the efficiency of
the slaking process. Adjusting an upstream measurement to a downstream value would result in
overshoot in this case.

The best solution is to choose your in-line measurement point wisely and locate the grab sample tap at
the same location.

THE IMPORTANCE OF TEMPERATURE

The first point to be clarified is: What role does the temperature compensated electrode play?
Electrodes with integral temperature elements only compensate for changes in electrode millivolt output
with temperature according to the Nernst equation. A temperature change of +/- 25 oC can result in a +/-
0.5 pH error or more if not properly addressed [2]. Temperature compensated electrodes do not
compensate for actual solution pH changes with temperature [4]. This change in solution pH with
temperature is real and should be measured.

Take, for example, an amine based etching solution used in circuit board manufacture. The actual pH of
the solution can shift nearly one pH unit as the etching bath heats up to process temperature. Allowing a
grab sample to cool before measuring it (incorrectly assuming that electrode temperature compensation
will account for the difference) will result in an erroneous correction on the process electrode. This will
result in a decrease in process efficiency in this application.

The ideal solution is to measure the grab sample at the process temperature. It is extremely important to
understand the impact of temperature if this cannot be accomplished. Many industrial settings gather
several samples and take them to a central measurement point. Temperature can vary significantly
during this time introducing disagreement between in-line and off-line measurements. A simple graph
Copyright 2005 by ISA - The Instrumentation, Systems and Automation Society.
Presented at 50th ISA Analysis Division Symposium; http://www.isa.org
of pH versus temperature can be very helpful when solution temperature dependence is high or non-
linear.

TIMING IS EVERYTHING

Process samples can change pH with time for various reasons. Carbon Dioxide uptake from air, for
example, can create an offset in pH (especially in solutions of low ionic strength). Biologically active
solutions can continue to react as microbes metabolize nutrients. Solutions can off-gas volatile elements
causing a pH shift especially in higher pressure processes.

Similar to the temperature effect discussed above, it may be beneficial to examine pH shift with time
experimentally so that this offset can be accounted for.

THE IMPACT OF ELECTRICAL GROUND

Ground loops are common in process applications particularly where high horsepower pumps or
galvanic processes are used. Offsets due to ground loops can occur because the reference element is
electrically connected to the process piping through the measured solution. The offsets can be small and
easy to overlook and are sometimes transient and therefore difficult to troubleshoot. Bear in mind that a
0.05 volt stray signal results in nearly one pH unit error. These grounding offsets do not impact the off-
line measurement because the sample is usually isolated from the process. This difference in grounding
conditions can create a discrepancy between the in-line and off-line measurement. Use of a “solution
ground” style process electrode helps to minimize these problems.

STORAGE OF THE PROCESS ELECTRODE

All electrodes are ideally stored per the manufacturer’s recommendations. There are times, in batch
processing for example, where the industrial in-line electrode will be stored in a relatively dry
environment. The active layer of the glass sensing element is actually a gel. This gel dehydration
creates a deviation from calibration. The longer the dry storage and the drier the environment, the
longer it will take an electrode to re-hydrate.

Once the electrode hydrates fully (both the glass and the reference junction) the electrode will usually
return to its calibration values. The pH reading will drift during hydration and there will be a true
differential to the off-line (correct) pH reading [1][3].

A complete and easy to automate solution to off-line storage is the retractable housing (Figures 4A and
4B). These housings allow the sensor to be retracted from the process vessel or line without disturbing
the process. Retractable housings are available either as manual, or pneumatically actuated designs.
Pneumatic housings are designed for easy automation for remote retraction, flushing and wet storage of
the electrode.

Copyright 2005 by ISA - The Instrumentation, Systems and Automation Society.

Presented at 50th ISA Analysis Division Symposium; http://www.isa.org
 STAGES:
#1 #2 #3 #4 #5
Housing

Cleaning
Port

Process
Pipe

FIGURE 4A – RETRACTABLE FIGURE 4B – RETRACTION STAGES

HOUSING

It should be noted that electrode storage is a source of deviation between in-line and off-line
measurements that may be addressed either by prevention via automation or tolerance for resulting error.

MEASUREMENT TECHNIQUE

Several different people are often involved in taking off-line readings in an industrial setting. The
technique used to calibrate, sample, and measure pH grab samples should be standardized to the point
there is no interpretation from user to user.

One question that arises in both calibration and measurement is whether to stir or not to stir. The simple
answer to this is: unless there is a method of continuously stirring e.g. a stir bar, do not stir the sample or
buffer. Make sure the impact of volatiles as discussed above have been addressed when stirring is used.
It is important to mix samples that tend to stratify to keep samples homogeneous.

Sampling technique is equally important. Flush the sample line to the point where a representative
process sample can be obtained. It cannot be expected to get a reading that will accurately compare to
the in-line electrode otherwise.

RECORD THE IN-LINE MEASUREMENT

Remember to record the value of the in-line measurement at the time the grab sample is taken. The pH
value is likely to have changed by the time the in-line analyzer reading is adjusted. This seems simple,
but in many industrial settings a sample is taken, analyzed in the lab and the dated value is applied to the
process, regardless of the time between sampling and reporting.

The solution is very basic: Record the in-line reading at the time of sampling and calculate the
differential versus the grab sample. Apply the differential to the in-line analyzer via a one-point
standardization if this error exceeds the acceptable tolerance.

Copyright 2005 by ISA - The Instrumentation, Systems and Automation Society.

Presented at 50th ISA Analysis Division Symposium; http://www.isa.org
 DEGRADATION OF THE PROCESS ELECTRODE

All of the errors addressed above can be avoided with relative ease. Most of them rely upon the user
having a good understanding of their process and their equipment. The topic of electrode failure
analysis and degradation within the process is wide in scope. There are some basic points, however, that
can be addressed here. First, heavy fouling, extreme temperatures or excessive pressure will degrade the
electrode slope and/or increase offset. Sometimes an in-line electrode can go to failure in a day. One
choice in these cases is frequent adjustments via comparison to grab samples. It is important that grab
sample frequency must be high enough to keep the process within error tolerances. Another choice is
implementing automation to minimize fouling.

There are many automation solutions that address most tough applications and in many cases eliminate
the need for grab sample adjustment [5]. These systems address many of the in-line concerns discussed
above: auto-calibration with no operator involvement (repeatability), automated wet lay-up (eliminating
hydration concerns) and automated cleaning (addressing degradation). Full automation significantly
decreases the required frequency for off-line measurements, resulting in labor savings in addition to
improved performance. Simpler automation systems will often address the majority of the degradation
issues. Heavy fouling, for example, can be addressed by a periodic “retract and flush” system at a
fraction of the price of full automation.

SYSTEM COMPONENTS:

x CONTROL UNIT
x VALVE ASSEMBLY
x TRANSMITTER
x DCS/PLC
x RETRACTABLE HOUSING
WITH SENSOR
x INDICATOR LIGHT
x BUFFER/REAGENT
CONTAINERS

FIGURE 5 – COMPLETE AUTOMATION SYSTEM PROVIDES SENSOR RETRACTION,

RINSE,CHEMICAL CLEANING,AND TWO-POINT CALIBRATION

Copyright 2005 by ISA - The Instrumentation, Systems and Automation Society.

Presented at 50th ISA Analysis Division Symposium; http://www.isa.org
 CONCLUSION

The above primer details most causes of offsets between the in-line measurements and grab samples.
Suggested solutions revolve primarily around understanding the process application and attention to
detail in calibration, sampling and measurement.

Automated systems are available that deliver many benefits as well as a sound return on investment
when electrode degradation is a concern or where full optimization is desired.

Understanding the limitations of pH measurement technology coupled with the above mentioned
optimization of errors will minimize off-line analysis requirements and deliver an in-line measurement
you can rely on.

REFERENCES

1. McMillan, Gregory K., Cameron, Robert A., Advanced pH Measurement and Control, 3rd Edition,
ISA - The Instrumentation, Systems, and Automation Society, Research Triangle Park, NC, (2005)
pp. 19, 21, 80.

2. Westcott, C. Clark, pH Measurements, Academic Press, New York, 1978, pp. 11, 102.

3. Bates, Roger G., determination of pH, Theory and Practice, John Wiley and Sons, Inc., New York,
1964, pp. 41, 335.

4. Dreyfus, Robert H., "D 1293-99, Standard Test Methods for pH of Water", 2002 Annual Book of
ASTM Standards, Vol. 11.01 Water (I), ASTM, West Conshohocken, PA, 2002, pp. 128-136.

5. Queeney, Kenneth M., “Automated pH Cleaning and Calibration Systems –A Modular Approach”,
ISA Technical Papers Collection, ISA Volume 454, P164, ISA - The Instrumentation, Systems, and
Automation Society, ISA EXPO 2004, October 5-7, 2004, Houston, TX.

Copyright 2005 by ISA - The Instrumentation, Systems and Automation Society.

Presented at 50th ISA Analysis Division Symposium; http://www.isa.org