Introduction to the Theory of

the Streaming Current Meter

Daniel Edney

Introduction
A streaming current (SC) meter is an instrument for measuring the charge that exists
on small, suspended particles in liquid. A streaming current meter (SCM) is the only
online instrument that can be used to measure coagulated particle stability for the
feedback control of coagulant dosage.
The streaming current meter (SCM) is also referred to as the streaming current
detector (SCD) or streaming potential detector (SPD) as well as streaming current
monitor and streaming current analyser (SCA) in various works. The word detector
is used in earlier literature instead of the words meter, monitor or analyser
because earlier SCMs were considered to only provide a qualitative indication of the
presence and sign of charge, rather than a useful quantitative measurement. The words
1: Meter, 2: Monitor and 3: Analyser all mean the same thing in this context,
except that the implication of sophistication increases in this order.
The difference between streaming potential and the more useful streaming current
will be explained in the following section.
This paper uses the more modern and descriptive term streaming current meter for all
of these instruments.

Colloidal Particles
Turbidity is caused by suspended particles in water in the size range of approximately
0.01 to 100 um in size. The larger fraction can easily be removed by settling. The
smaller particles, with sizes of less than 5 um are referred to as colloidal particles (or
colloids) and have extremely slow settling velocities and so cannot be practically
removed by settling. The behaviour of colloidal particles in water is strongly
influenced by their electrostatic charge. This colloidal charge comes about because of
the uneven surface characteristics of the particles and in most solids is negative,
particularly the alumino-silicate clays typically suspended in surface water. The
charge on each particle will repel others and prevent significant flocculation from
occurring. Neutralising this charge is the main purpose of coagulation.

However, simply considering the charge as an electrostatic surface charge is an over

simplification. As the whole solution must have a neutral charge, a layer of water
containing ions of opposite charge surrounds each colloid. Consideration of colloidal
charge purely in terms of the colloidal particles by themselves has no benefit, as the
surrounding ions are, for all practical purposes, inseparable from them.
The Double Layer
The double layer model is used to explain the distribution of ions around each
colloidal particle. This is a long standing conventional approach to colloid analysis, a
more detailed description is available in a number of works on this subject (eg Bratby
1980).
For illustrative purposes we will assume a negatively charge colloid. Closest to the
negative surface of the colloid there is a layer of strongly bound positive ions this is
know as the Stern layer. Further positive ions are still attracted to the colloid but are
repelled by the Stern layer; likewise, negative ions are attracted to the positive ions
but repelled by the colloid. A dynamic equilibrium of negative and positive ions
forms outside the Stern layer, known as the diffuse layer. The concentration of
positive ions in the diffuse layer gradually decreases as the distance from the colloid
increases until beyond a certain distance the ion concentrations are the same as the
equilibrium in the water.
The strongly held positive ions near the surface and the charged layer surrounding this
is where the name double layer comes from. The point just outside the stern layer is
often called the shear plane as motion of the particle through the water shears the
diffuse layer away.
The thickness of these layers depends on the concentration of ions in solution. At any
distance from the surface the charge density is equal to the difference between the
concentration of positive and negative ions at that distance. This results in an
electrical potential (which is a voltage, caused by separation of charges) existing
across the layers. This potential is greatest near the surface and decreases to zero as
the distance from the colloid increases. A graph of this potential curve is useful
because it indicates the distance at which the interaction between colloids will occur.
This is illustrated in Figure 1.
The potential at the boundary between the Stern layer and diffuse layer is called the
zeta potential. Zeta-potential is useful, as it is a direct indication of the amount of
energy required to bring separate particles together.

Figure 1 Ways to visualise the Double Layer model and the potential as a
function of distance from the colloids charged surface.
In fresh water suitable for treatment, the diffuse layer is large compared to the stern
layer and so zeta potential gives a good indication of surface potential, which is
directly related to surface charge. Zeta potential can be measured quite easily. This is
done by using a microscope to observe turbidity particles inside a thin chamber called
an electrophoresis cell. An electric field is applied along the cell, if the particles are
negatively charged they will move towards the positive end of the cell, carrying their
stern layer ions with them. The average speed with which these particles move
depends on the zeta-potential, as it is this that determines the net electrical force
operating on the particle and its stern layer ions.
In practice, the relationship between zeta-potential and this average speed is not
always straightforward to calculate, as it is dependant on the fluids viscosity,
dielectric constant, conductivity and temperature. As a result, zeta potential, which
should have units of mV, is often expressed in terms of electrophoretic mobility. This

has rather confusing units in velocity per electric field strength, typically um/s per
V/cm.
Other disadvantages of measuring zeta-potential in this way include:

It can only be determined for particles that are large enough to be detected and
tracked through a microscope.

It requires a lab technician to operate the apparatus and to observe the particles
through a microscope. Therefore, Zeta potential measurement is slow, labour
intensive and cannot be used online.

It is not very accurate near zero, which unfortunately is the point of greatest
interest, because of the difficulty in tracking the particle motion.

Zeta-potential measurement has applications in laboratory studies of coagulation and

flocculation processes, but is of little use for online automation.

The Streaming Current Effect

The streaming current (or streaming potential) effect is a reverse manifestation of the
electrophoresis effect. This occurs when a particle is mechanically moved through the
fluid, or the fluid moved past the particle, resulting in a separation of charges causing
a potential to exist.
A commonly situation occurring situation is that water with negatively charged
particles is forced through a filter. The negatively charged particles become lodged in
the filter, while the mobile positively charged ions are swept downstream. This
separation of charges causes an electrical potential to exist across the filter, called the
streaming potential. If electrodes are inserted upstream and downstream of the filter
then the electrical potential can be measured. The upstream electrode is negative and
the downstream electrode is positive. (Electrical potential is also referred to as
voltage, as it is measured in unit called Volts.)
The electrical potential results in a current being conducted upstream through the
water to remove the charge separation, this is the streaming current. The electrical
potential therefore depends on the conductivity of the liquid. If the electrodes are
connected together by a path with a much higher conductivity than the water, such as
a wire, then the current will flow mostly through this path and little potential will
exist.
Streaming current is a more fundamental quantity than streaming potential, as it does
not depend on the waters conductivity.
Streaming current can be measured whenever the water with suspended charged
particles is forced through thin capillaries or other barriers to the particles motion.
However, useful measurements are extremely difficult to directly make from this
effect, as the current produced is extremely small and easily obscured by potentials
that exist for other reasons and electrical paths through the environment. Difficulties
such as the dc offset in the electronics and electrode drift make this simple approach
virtually impossible to use in a practical sensor.

Strictly speaking, zeta-potential is a property of a surface, rather than of a group of

particles. A variety of surfaces can produce zeta-potential (and streaming current) in
water. These, normally insignificant, effects are well recognised in materials science.
This work, however, uses those terms in the manner typical of discussion focused on
water treatment, and considers them to be bulk properties of suspended colloids. Thus
a sample of turbid water is said to have a certain streaming current, when correctly it
is the average zeta-potential of the surfaces of the suspended colloidal particles that
produces a certain streaming current in a certain sensor1.

A Practical SCM Sensor Design

The practical streaming current meter (SCM) was invented around 1966 and credited
to F.W Gerdes. Gerdes approach overcomes many of the limitations of direct
streaming current measurement (Gerdes 1966 a and b) and is the basis for all modern
SCMs.
The SCM is based on the effect where the walls of the capillaries through which the
colloidal material flows quickly gain a coating of particles and take on the surface
charge characteristics of these particles. The SC sensor consists of a piston and a
close-ended chamber. A narrow gap, an annulus 200-500um wide, exists between the
piston and the walls of the chamber The piston is driven up and down at a fixed
frequency, typically 4-5 strokes per second, forcing sample water in and out of the
chamber through the annulus. A typical SC sensor is shown in Figure 2.

Figure 2 Streaming current sensor.

It is clear why the former terminology is preferable.

As the piston and chamber surfaces are coated with charged particles, the water
flowing rapidly up and down through the annulus results in displacement of the
counter-ions. The SC signal measured by electrodes in the annulus is proportional to
the water velocity and therefore alternates in time with the piston. This signal is
typically in the range of 0.05uA to 5uA depending on the particular conditions.
Measurement of SC in a closed chamber has several advantages compared to
measurement directly in a flowing stream:

The closed end is electrically isolated and removes problems caused by large
potentials in the process stream from other sources.

The signal is alternating at the frequency of the piston. This allows it to be

separated from external noise and offset caused by electrode drift and
dissymmetry.

Practical aspects of instrumentation, such as that the closed chamber can be

shielded from electromagnetic interference and cleaned easily.

Theoretical Description
Unfortunately, a complete, quantitative and verified mathematical description of the
sensor does not exist.
Several attempts have been made at modelling the SC sensor. The original approach
by Gerdes (1966a) was based on a simplified triangular fluid velocity profile within
the annulus. Later work (Elicker et al 1992) used a more appropriate fluid profile
model. The most complete work to date also considered inertial effects within the
annulus and compared the solutions with those found by earlier methods (Walker et
al. 1996). They found that for the geometries used by typical SCMs, the earlier
approximate solutions produce very similar results to their more complete model.
All of these models have the following form:
I = k.s....f(r,R)
Where I=average current magnitude, s=piston stroke length, =motor cycles per
second, =dielectric constant of solution, =zeta potential, r= piston radius,
R=chamber radius, k=electronics gain constant, f()=a function of the annulus shape
which depends on the model used.
Experimental results show that the linear relationships with s and exist as predicted
(and, of course, k). Likewise, results have shown that I is generally linearly related to
. However the models predictions of the actual I measured are generally poor.
All the models assume that the surface of the piston is completely and uniformly
coated by the colloidal particles and by any coagulant in solution in a representative
manner. This is not necessarily a valid assumption and probably goes a long way
towards explaining the poor predictive abilities of this theory (Barron et al. 1994;
Dentel and Kingery 1989; Dentel et al. 1989a). Gerdes acknowledged the poor
predictive abilities of this type of model, and acknowledged the existence of double-

equilibrium theory. This states that the charge on a particle surface is the result of an
equilibrium between the ions on the surface and in the solution. The charge on the
piston surface depends on another but similar equilibrium between the same
concentrations in the liquid and a different adsorbed density on the piston surface. The
charge on the two different types of surfaces, which are both in equilibrium with the
same liquid, will be related to one another in a way that depends on their respective
equilibrium constants. Using this theory, it is only possible to predict that any change
in the colloidal system that changes zeta-potential will change SC in the same
direction.
The modern colloidal-science theory works implicitly reject this equilibrium-based
theory and provide no discussion of how, or at what rate, the measurement surfaces
take on the zeta-potential characteristics of the sample. If an equilibrium does exist
between the charge on the particles in solution and the charge on the measurement
surfaces then it would influence the rate at which measured SC responds to changes in
coagulant dosage. Indeed it can be shown that by starting with this assumption it is
possible to develop a simple dynamic model that is a good fit for the sensors
observed dynamic response (Edney 2005).
Even if a perfect theoretical model of the SC sensor was possible and did exist, it
would still be of limited practical use. This is because it would require complete
characterisation of the piston surface and the precise geometry of the annulus. Surface
characteristics and the tight tolerances are easily altered by physical wear and tear,
even over relatively short time frames. It is quite possible to calibrate an SCM so that
it reads directly in units of zeta-potential over a limited range or in units matching
another SCM (Elicker et al 1992; Dentel et al 1989a), however it will not stay
calibrated like this when in continuous use as the condition of the surfaces changes
with time.
The unit of SC should properly be the unit of current (the Ampere). However, the lack
of a consistent relationship between this and any useful physical or chemical
parameters means that SCMs normally use arbitrary units. Their scales are designed
to be adjusted for convenience and can be arbitrarily set by their users. This thesis
will follow this convention and use the term SC units to refer to arbitrary units used
to measure SC.

Usage and History

The most useful way to apply a SCM in a typical water treatment plant is for the
feedback control of coagulant dosage. The sample is taken from after the rapid mixing
stage and the SC continuously measured so that a controller can adjust the coagulant
dosing rate to maintain a set SC. This is illustrated in Figure 3. The SC that
corresponds to the desired plant operation is traditionally determined by manual jar
tests.
Figure 4 shows a typical relationship between alum dosage, SC of the dosed water
and residual turbidity from the clarifier or jar tests. The SC that corresponds to
minimum residual turbidity is typically offset slightly from zero, because of effects
other than charge neutralisation. Generally a residual turbidity that is slightly less than
the minimum is considered acceptable in order to reduce coagulant usage.

As the raw water turbidity changes, the coagulant to residual turbidity graph shifts
considerably, while the relationship between SC and residual turbidity stays much
more constant. Controlling the coagulant dose such that the SC is constant at the
desired set point will result in a much more consistent residual turbidity. Often
coagulant chemical savings can be achieved because conservative over dosing under
manual control is avoided.

Figure 3 Diagram of SCM usage in a typical conventional plant.

5

0
Desired SC SP

-5

-10
1

10

15

20

25

30

35

40

0.8
0.6
Desired Turbidity

0.4
Minimum Turbidity Point

0.2

10

15

20

25
30
Alum dose (ppm)

35

40

Figure 4 Typical relationship between alum, SC and settled water turbidity.

Caution and a lack of understanding kept the SCM from being used in most plants
until the early 1980s. The study most widely cited, and credited with bringing
widespread acceptance for use of the SCM in water treatment is the work of Dentel
and Kingery (for a summary, see Dentel and Kingery 1989). This includes the result
that SCM use resulted in generally improving water quality and consistency, as well
as decreased chemical usage by 12% during stable operation and 23% during
changing raw water conditions. A survey of plants using an SCM showing 80% of
them regarded it as useful. It is likely that these percentages would be higher today,
even though many more SCMs are in use, due to better understanding by both users
and manufacturers. One of the most common problems experienced is clogging of
sample lines due to floc build-up. This can normally be minimised by use of an
appropriate sample point and sampling design.
Another significant issue that has been widely identified since then is the effects of
wear and tear on the measurement surfaces and the mechanisms. Any unevenness in
the stroke due to misalignment or looseness in the mechanical drive will translate
directly into a distortion of the SC signal. Changes in the condition of the sensing
surfaces will change the manner in which they become coated with colloidal material
and the equilibrium constants between the surfaces and the solution. These effects are
normally the most significant on the surface of the piston. Once the wear reaches a
certain level the reading produced becomes unstable, and commonly drifts in an
unpredictable manner. It is not always possible to tell by inspection if a particular unit
is in this condition.
The successful operation of a SCM has developed a (in the authors opinion,
unjustified) reputation among some WTP operators for being more of an art than a
science. The reasons include the difficulty in detecting failure, the arbitrary nature of
the measured units and the inconsistency between different plants and different
instruments.
However, the instruments usefulness is no longer in question. For example, the New
Zealand Ministry of Health requires plants supplying more than 5000 people to have
automatic control of their coagulant dosage in order to achieve either of the highest 2
grades, required to be graded more than merely satisfactory (Public Health Grading
Drinking Water Supplies 2003). Currently, a SCM is the only technology that can
provide this.

References
American Water Works Association. (1990) Water Quality and Treatment : a
Handbook of Community Water Supplies. 4th Edition. McGraw-Hill, New York. pp.
269-284.
American Water Works Association and American Society of Civil Engineers (1998)
Water Treatment Plant Design. 3rd Edition. McGraw-Hill, New York. pp. 87-110.
Barron, W, Murray, B.S., Scales, P.J., Healy, T.W., Dixon, D.R. and Pascoe, M.
(1994) The streaming current detector: A comparison with conventional electrokinetic
techniques. Colloids and Surfaces A: Physicochemical and Engineering Aspects, 88
(19), 129-139.

Bellamy, W.D., Cleasby, J.L., Logsdon, G.S., and Allen, M.J. (1993) Assessing
Treatment Plant Performance. Journal of the American Water Works Association,
85(12) 3438.
Bratby, J. (1980) Coagulation and flocculation : with an Emphasis on Water and
WastewaterTreatment. Uplands Press, Croydon.
Dentel, S. K. and Kingery, K. M. (1989) Using Streaming Current Detectors in Water
Treatment. Journal American Water Works Association, 81, 85-94.
Dentel, S. K., Thomas, A. V., and Kingery, K. M. (1989) Evaluation of the Streaming
Current Detector I. Use in Jar Tests. Water Research, 23, 413-421.
Dentel, S. K., Thomas, A. V., and Kingery, K. M. (1989) Evaluation of the Streaming
Current Detector II. Continuous Flow Tests. Water Research, 23, 423-430.
Edney, D.B.L (2005) Control and Optimisation of Coagulant Dosing in Drinking
Water Treatment. PhD Thesis. University of Auckland.
Elicker, M.L., Resta, J.J., Hunt, J,W. and Dentel, S.K. (1992 ) A Fundamental Basis
for Use of the Streaming Current Detector, Proc. 1992 AWWA Annual Conf.,
(Vancouver). Jun 22.
Gerdes, F.W. (1966) A New Instrument The Streaming Current Detector. Analysis
Instrumentation, 4, 181-198.
Gerdes, F.W. (1966) Apparatus for Measuring Charge Condition Within a Solution.
US Patent number 3 368 145.
Public Health Grading of Community Drinking- Water Supplies 2003 - Explanatory
Notes and Grading Forms (2003). New Zealand Ministry of Health. pp 34-41.
Walker, C.A., Kirby, J.T. and Dentel, S.K. (1996) The Streaming Current Detector: A
Quantitative Model. Journal Of Colloid And Interface Science, 182, 7181. (Online:
http://chinacat.coastal.udel.edu/~kirby/papers/walker-kirby-dentel-jcis96.pdf)