Proceedings of the 9th WSEAS International Conference on Automatic Control, Modeling & Simulation, Istanbul, Turkey, May 27-29,

2007 58

Bath and Deposit Monitoring System for Electroless Nickel Plating Process
ROBERT TENNO, KALLE KANTOLA
Department of Automation and Systems Engineering
Helsinki University of Technology
Otaniementie 17, 02015 TKK
FINLAND
robert.ten http://www.control.tkk.fi

Abstract: - Electroless nickel plating is a widely used plating method in many industries. The usability of the coating is,
however, restricted by the control problems related to the electroless nickel plating process. To ensure deposit quality, the
transcendent properties of electroless nickel deposit should be accurately controlled during plating, which is challenging
because of the spontaneous, unstable reaction and lack of online measurements. To improve operator based control, a special
monitoring system for electroless nickel plating is presented in this paper. The monitoring system is calibrated especially to
the electroless nickel process used for Plated-Through-Hole (PTH) board manufacturing. In the paper it is shown, that, the
system is capable to estimate the electrical, chemical and deposit parameters of the plating process from standard
measurements of nickel plating industry. The proposed monitoring system enables also the replace of the individual
parameter supervision by the supervision of the overall process dynamics, i.e. the electrochemical processes.

Key-Words: - Monitoring, state estimation, electroless plating.

1 Introduction controlled parameters and because of the fact, that the exact
Electroless nickel plating is one of the amorphous materials reaction mechanism of the plating reaction is not known
used in industry [1]. It is a widely used plating method in [1,7,8].
many industries, covering applications from oil and gas In this paper, a new model based monitoring system
industry, machinery, etc. to electronic and micromechanic for the electroless nickel plating process is proposed. The
manufacturing [1]. The broad applications variety of the monitoring system calibrated and tested especially to the
plating is because of the advances of electroless nickel coat- electroless nickel process used for Plated-Through-Hole
ing, including uniform thickness distribution, high wear (PTH) board manufacturing, because the challenges of the
resistance and dense morphology [1,2]. These properties electroless nickel plating are emphasized by the tight
makes the electroless nickel alloy an excellent protection tolerances and challenging, detailed geometries of the PTH
against oxidation and corrosion [3,4], which is its main use. industry. The proposed monitoring system can, however, be
If compared to electroplated nickel, the uniform thickness applied also to other applications as well. The electroless
distribution of the electroless nickel alloy enables the nickel plating is an established process in manufacturing of
plating of more detailed and challenging geometries [3]. printed circuit boards, where electroless nickel alloy works
In addition to these mechanical properties, the as a diffusion barrier between coppered circuitry and golden
electroless nickel alloy is lead-free, which makes it a very surface layer in so called electroless nickel immersion gold
usable for electronic applications, like in printed circuit (ENIG) surface finish [1]. In this technology, the constricted
boards (PCB), in which the lead is tried to be replaced by spaces in conducting holes are plated with a thin nickel film
other materials. In electrical devices sold in the markets of containing predefined percentage of phosphorous ensuring a
European Union, the use of lead is even prohibited by stress-free, dense morphology. Unfortunately, the most
legislation after 1.7. 2006 [5]. Therefore, it is likely, that the critical deposit parameters i.e. film thickness and
electroless nickel plating will be more popular in the future. phosphorous content cannot be measured online during
The usability of the coating is, however, restricted plating. Instead, these parameters are measured indirectly
by the control problems related to the unstable and after plating by X-ray fluorescence (XRF) based coating
spontaneous electroless nickel plating process. These thickness gauge (CMI) analysis and wetting balance test
problems include, inter alia, the plating bath contamination with approximately 10 hours delay. Because of the long
[6], meaning the precipitation of the bath caused by the delay, these measurements are not suitable for effective
unbalanced and too aggressive reaction. In addition, the co- process monitoring or control. Also, the possible errors in
deposited phosphorous has a critical effect to the the deposit cannot be corrected afterwards and therefore, it
morphology and further to the properties of the deposit [1]. is essential to monitor the plating reaction online during the
The control of the plating reaction is a challenging task, plating to ensure a uniform quality of the deposit.
because of the lack of online measurements of the
 Proceedings of the 9th WSEAS International Conference on Automatic Control, Modeling & Simulation, Istanbul, Turkey, May 27-29, 2007 59

The proposed monitoring system provides resistances [1,3]. Therefore a monitoring method is needed
essentially more information of the process online than the to determine the optimal plating conditions, related to the
traditional methods provide offline. Besides the alloy process state.
thickness and phosphorous content, the system provides the
bath parameters in terms of the electrical and chemical
characteristics, which can be exploited to failure source 3 Model
detection and process control. The monitoring system is The discharge model [9,10] of the reactions (1)-(4) is a two-
based on a model developed earlier by the authors [9,10]. directional electrode model where the current densities of
This model has been validated before and therefore, not the the reactions are calculated from the Buttler-Volmer
model itself but only the monitoring system is analysed in equation and controlled through empirical formulas
this paper. The purpose of the paper is practical – to accelerating or decelerating the reactions according to the
introduce an industrial application for the monitoring of concentration of reagents. The equilibrium potentials of the
electroless nickel plating process. reactions (1)-(4) are calculated from the Nernst equation
and a mixed potential from the current balance of the anodic
and cathodic reactions. The concentration dynamics of the
2 Reaction reagents is evaluated from the mass balance.
The hypophosphate reduced electroless nickel plating is a The model enables online estimation of the product
complex and challenging process. As a testimony of that, its parameters (film thickness, phosphorous weight
reaction mechanism is not fully known but several percentage), along with the electrochemical parameters
mechanisms are proposed for the reactions [1,7,8]. At the (deposition speed, mixed and equilibrium potentials, current
moment, the electrochemical reaction mechanism is the densities) and the chemical parameters (reaction rates,
most seconded one [1] and it is used by the authors as a hypo- and orthophosphate density, hydrogen evolution rate)
base for the developed electroless nickel plating process from the measured pH-index, nickel concentration, reagent
model [9,10]. This model is applied for process monitoring addition flow rates, bath temperature, plating time and bath
in this paper. The model is based on the following anodic loading. The data flow sheet is shown in Fig. 1. It represents
and cathodic reactions. a complex calculation engine (model calculation), which is
not discussed in this paper but detailed description can be
Anodic reaction – hypophosphite oxidation found from Tenno et al. [10].
This model-based monitoring provides the current
H2PO2- + H2O R H2PO3- + 2H+ + 2e-. (1) state of the product (board) and bath parameters faster and
more completely than in any laboratory analyses. In the case
Cathodic reactions – phosphorous deposition, hydrogen gas of a failure, it allows fast detection of the failure source,
evolution, nickel deposition which is essential for failure correction. This monitoring
system is discussed next.
H2PO2- + 2H+ + e- R P + 2H2O, (2)
Controller Product

2H + 2e R H2,
+ -
(3) pH-index Film thickness

Phosphorous
Ni2+ + 2e- R Ni. (4) Nickel
concentration content

In these reactions, the hydrogen ion production exceeds Flow rates Bath
Model

consumption and the degreasing of pH-index is balanced Deposition speed

with ammonia additions, when the following neutralizing
reaction takes a place Temperature Potentials, current
densities
NH3 + H3O+ NH4+ R + H2O. (5) Plating time
Reaction rates

In practice, the pH-index is increased depending on the Hypo and orto-

Bath loading
metal turn over (MTO) to compensate the decreased phosphate density
reaction rates affected by the bath aging i.e. accumulation of
the by-products like orthophosphate into the bath. Fig. 1. Model-based process monitoring.
Generally, the increased pH-index decreases deposit’s
phosphorous content degrading board’s corrosion
 Proceedings of the 9th WSEAS International Conference on Automatic Control, Modeling & Simulation, Istanbul, Turkey, May 27-29, 2007 60

4 Monitoring System
The process monitor shown in Fig. 1 was integrated to the 5 Monitoring Validation
control system existing at a PTH factory. It is in interaction The process monitoring system, shown in Fig. 2, is based on
with the electroless nickel plating bath through the bath a process model which was originally calibrated on a large
controller and flightbar controller as shown in Fig. 2. For sample of industrial data and shown to be accurate [10].
time series analysis option, it is also in interaction with the Because of the model dependence, the monitoring systems
laboratory database. Each element of the system has accuracy is directly related to the accuracy of the process
multiple functionalities described in the following. model. In this paper, the monitoring system is tested
retrospectively on a large sample of independent industrial
data shown below. This data represents the plating bath
Plating bath behavior during 6 bath lifetimes, which is equivalent to
pH control Bath loading
Ni control Temperature 284,390 PTH boards made. The monitoring validation is
control
Plate Solution done by comparing the monitored values on the measured
probe probe ones for the processes in which the measurements were
Pump
calibration
available. This is done in Figs 3, 4, 5, where concentrations
Bath Laboratory Flightbar
controller database controller
are compared to the laboratory analyses.
4.9
pH-index Measured
Film thickness Ni density Target
4.85
Model
Solderability Hyp. density Lab
pH-index Temperature
4.8
Ni density Plating time
Flow rates Process Loading events
4.75
monitor pH

4.7
Fig. 2. The monitoring system in which the monitoring
4.65
engine of Fig. 1 is integrated with the plating bath through
the intermediate layer. 4.6

4.55
Bath controller This controller keeps the plating
4.5
chemistry in balance by adding nickel sulphate, sodium 0 5 10 15
Process time, days
hypophosphite and stabilisers in the bath. It also controls the
pH-index according to the target profile. Fig. 3. The measured, model-predicted and laboratory-
Flightbar controller This is a transport crane, analysed pH-index.
which immerses or elevates the PTH billets into/from the 5

bath. It also manages the logging register and takes care of 4.9
the plating temperature, performing measurements and 4.8
heating. 4.7
Laboratory database The laboratory analyses 4.6
Nickel, g/dm3

carries out the final quality checking and supervises

4.5
operators if some correction is required. The analyses
4.4
include the product parameters (film thickness, wetting time
4.3
and force), the bath parameters (pH-index, nickel and
hypophosphite concentrations) and the calibration of the 4.2 Measured
Target
peristaltic pumps of the controller if some systematic 4.1 Model
Lab
deviation between online measurements and laboratory 4
0 5 10 15
analysis is discovered. Process time, days

Process monitor This computer characterises the Fig. 4. The measured, model-predicted and laboratory-
current state of the process in terms of electrical, chemical analysed nickel concnetration.
and board parameters (see Fig. 1). It can also represent the
plating history up to the current moment and compare these The monitoring model estimated and the measured
with the laboratory analysis. The process monitor pH-indexes are in relatively good agreement with each other
downloads the latest data from the bath controller, flightbar and in moderate agreement with the results measured later
controller and from the laboratory database. in laboratory analysis as can be seen in Fig. 3. The same
Currently the monitoring system is under testing in agreement is relevant to the nickel concentration shown in
the industrial plating process. One of the test runs is Fig. 4. However, there exists some error in nickel
discussed next. concentration estimate, which might be because of the error
 Proceedings of the 9th WSEAS International Conference on Automatic Control, Modeling & Simulation, Istanbul, Turkey, May 27-29, 2007 61

in the flow measurements of the reagent additions. Despite 5.5

this, the estimated nickel concentration seems to be closer to Alloy
Target
the laboratory measurements than the bath controller’s ones. CMI
5
As discussed above, the bath controller adds nickel sulphate
into bath to replace the nickel ions consumed by the

Thickness, um
4.5
reaction. The target is to keep the nickel concentration
constant, which seems not be the case as can be seen from
the figure, where the nickel concentration changes 4

vigorously. These changes are because of the changes in the

bath loading, which cannot be avoided in an industrial 3.5

plating process. The proposed monitoring system might

help to cancel the effect of the loading, because it takes the 3
0 5 10 15
loading into account beforehand and therefore reacts more Process time, days

rapidly to the concentration changes that the analyzer of the Fig. 6. The film thickness; estimated online and later
bath controller. measured in CMI analysis.
50
Target
Model other and the trends of the measurements and estimation are
Lab
45 the same. There are, however, some occasional exceptions,
which might be caused by bath start ups or exceptional bath
Hypophosphite, g/dm3

40
loadings. The peak values of thickness may be explained
with the corresponding peaks of the measured nickel
35
concentration (Fig. 4) and with the start-up concentrations
30
of new baths (Figs 3, 7). The fast oscillation of the value is
because of the corresponding changes in plating time and
25 plating area. If the film thickness is compared with the
measured pH-index (Fig. 3) or MTO (Fig 7), these seem to
20
0 5 10 15
be related. This relationship is natural, because it is ob-
Process time, days
served [1,2], that higher pH-index accelerates plating
Fig. 5. The online-estimated hypophosphite concentration, reaction while high MTO for its part decelerates plating
compared to the results measured later in a labo-ratory. rates. To guarantee stable product quality in PTH
manufacturing, the plating rates are desired to keep constant
The estimated and laboratory measures and therefore, the decelerating affected by the MTO raise is
hypophosphate concentrations are presented in Fig. 5. As compensated by increasing pH-index along with MTO.
can be seen, the hypophosphate estimate follows the Currently, the target profile of pH-index seems to depend on
laboratory measurement and the estimation accuracy is the MTO-index as a linear function implemented through
related to the nickel estimation (Fig. 4). The hypophosphate small steps as show in Fig. 3. This is not accurate enough
works as a reducer in the studied process and its and the target profile should be chosen more precisely if
consumption is assumed to be related to nickel uniform quality is desired, as can be seen from the figure.
consumption. Therefore, its concentration is not measured This means that instead of linear stepwise relationship, the
by the bath controller, but the hypophosphate is added along target profile should be continuously dependent on MTO
with nickel addition. In a well balanced process, a and also reagent concentrations, like on nickel and
hypophosphite concentration should be in balance with the hypophosphate concentrations.
nickel concentration, which seems to be the case if In general, the phosphorous content depends on the
comparing the Figs. 4, 5. However, there exists some pH and MTO-indexes [1]. In Fig. 6, the estimate of the
variation between the measurements. This variation is phosphorous weight percentage, characterising the tensile
mainly because the fact that nickel takes a part only in strength and corrosion resistance of the film [1,7,8,12], is
reaction (4) while hypophosphate is involved in reactions shown. As in the case of film thickness, the peak values of
(1) and (2). Therefore, the used control method is not phosphorous may be explained with the corresponding
accurate, but the hypophosphate should be controlled peaks of the measured nickel concentration in Fig. 7 and
individually, which can be done by using the monitoring somewhat less with the start-up values of a new bath (Figs
system. This would improve the bath balance and prevent 3, 7). The phosphorous content cannot be measured online
the bath contamination affected by unbalanced bath [11]. and therefore no phosphorous measurement is presented as
In Fig. 6, the online estimated and CMI-measured a validation reference. Phosphorous content is, however,
film thickness (product parameter) is shown. The estimated closely related to solderability of PTH [13,14] and
and measured values seem to be relatively close to each
 Proceedings of the 9th WSEAS International Conference on Automatic Control, Modeling & Simulation, Istanbul, Turkey, May 27-29, 2007 62

therefore, it can be estimated from solderability speeds. The observed peak values in the deposition speeds
measurements presented in the figure. are because of the corresponding peaks of the current
200 4
densities (Fig. 10). The current densities are calculated from
H2PO3
MTO
reagent concentrations through Buttler-Wolmer equation
and the peaks in the current densities are, for their part,
150 3 affected by the changes in reagent concentrations. The
shapes of the deposition speeds and current densities curves
Ortophosphite, g/dm3

are inherited from the equilibrium potential curves, shown

MTO
100 2
in Fig. 11.
9
Ni+P
50 1 Ni
8
P

Deposition speed, um/h

0 0
0 5 10 15 6
Process time, days
5
Fig. 7. The measured metal turn over and online-estimated
orthophosphate concentration. 4

In solderability measurements the time T is the time

2
when the solder reaches 2/3 of its maximum force and the
wetting forces F1 and F2 are the maximum forces of the 1
0 5 10 15
solder 2 and 4 second after soldering. Although, some Process time, days

correlation can be seen between the estimated phosphorous Fig. 8. The online-estimated deposition speed total for alloy
content and the solderability results (the phosphorous and volumetric speed for nickel and phospho-rous.
content correlates with the wetting time with a correlation 10 3
Ni+P
coefficient = 0.14 and has almost no correlation (correlation Target
F1
coefficients -0.028, -0.023) with the wetting force), the F2
T
estimation seems not be very accurate as can be seen from

Wetting time and force, sec, mN

9.5

the figure. This might be, because of the solderability

Phosphorous, wt %

characteristics are affected not only by phosphorous content

9 1.5
but also by many other factors like the cleanliness of the
surface, activation of the surface, board aging at high
temperature, nickel concentration in gold bath and gold 8.5
layer thickness of ENIG surface. Therefore, it seemed that
these characteristics alone are poor indicators for the
phosphorus content of the film. The phosphorous content is, 8
0 5 10
0
15
however, a critical parameter for corrosion resistance and Process time, days

therefore it should be supervised online. For that purpose, Fig. 9. The phosphorous weight percentage; estimated
the proposed monitoring system seems to offer an effective online and later measured wetting time (T) and forces (F1,
tool. F2).

The overpotentials i.e. the voltages between the mixed

6 Process Dynamic Analysis potential and equilibrium potentials (see Fig. 11), are inputs
Although, the reaction mechanism of the electroless nickel for the Buttler-Wolmer equation and, therefore, they control
plating is not completely known, the complex relationships partly the current densities of the partial reactions (Fig. 9).
between the partial reactions of the process can be analyzed Through Buttler-Wolmer equation, they affect on the
by using the online-estimates, provided by the monitoring deposition speed of nickel and phosphorous (Fig. 8) and
system as discussed in the following. eventually on the film thickness (Fig. 6) and phosphorous
In Fig. 8, the deposition speed of the deposit is content (Fig. 9). Moreover, the over potentials and current
decomposed to speeds of nickel and phosphorous densities are related also to the chemical reaction rates show
depositions. While the total speed controls the film in Fig. 12. The chemical reaction rates determine the
thickness (Fig. 6), the balance of its components controls concentration dynamics of the reagents and are also affected
the phosphorous content of the alloy (Fig. 9). This is by the reagent concentrations itself. Therefore, direct
because the deposited masses are integrals of the deposition relations between the overpotentials, current densities and
 Proceedings of the 9th WSEAS International Conference on Automatic Control, Modeling & Simulation, Istanbul, Turkey, May 27-29, 2007 63

the reaction rates are rather hard to see by comparing the 150 15
respective figures. H2PO2
H2
Ni
P
100 10

Reaction rate: H2PO2, g/min

40 8

H2, Ni, 10xP, g/min

50 5
Current density: i3, i1, mA/cm2

i1
20 i3 4

i4, 5xi2, mA/cm2

i2 0 0
i4

0 0
-50 -5

-20 -4
-100 -10
0 5 10 15
Process time, days

-40
0 5 10
-8
15 Fig. 12. The online-estimated reaction rates of
Process time, days
hypophosphite oxidation (H2PO2), hydrogen gas evolution
Fig. 10. The online-estimated current densities of (H2), nickel (Ni) and phosphorous (P) deposition reactions.
hypophosphite oxidation (1), phosphorous deposition (2), 0.35
hydrogen gas evolution (3) and nickel deposition (4)
reactions. 0.3

0.25
In the reaction (3) hydrogen gas is formed from the
hydrogen ions. An increase in gas formation is observed to g/dm3/min
0.2

precede the bath contamination by Winowling Jappes et al.

0.15
[11] and therefore the estimate is important in process
supervision. As can be assumed, the direct measurement of 0.1

hydrogen gas evolution is difficult to carry out in an

0.05
industrial process environment making the monitoring
concept desirable. During the test, the hydrogen gas 0
0 5 10 15
formation is rather unstable as shown in Fig. 13 where the Process time, days

graph is a half hour average. The formation rate is related to Fig. 13. The online-estimated hydrogen gas evolution.
reaction rates which is natural: the hydrogen gas is a by
product of the deposition reactions. In the case of unstable As discussed, the concentration of species (pH-index,
hydrogen evolution the phosphorous content between the nickel, hypo- and orthophosphate) controls mainly the
film layers is hardly unique and a layered morphology or current densities, through Buttler-Wolmer equation, which
sometimes “black nickel” can be formed. The latter is furthermore controls reaction rates. Therefore, in electroless
closely related to weak solderability and is a problem in plating the concentrations are used for process control and
PTH technology [8]. especially the pH-index (Fig. 3) is used to compensate the
reaction rate decrease caused by the bath aging. The bath
-0.3
aging characterised by the MTO-index is mostly due to the
orthophosphate accumulation which might be a better
-0.4
indicator than MTO. Both of these are presented in Fig. 7.
-0.5 By comparing the figures 3 and 4, one can see that, on a
Potential, V

contrary to pH-index, the nickel concentration is kept

-0.6
constant during plating. This control strategy is in
-0.7 disagreement with the stable thickness and phosphorous
-0.8 U1 content requirement [6]. The phosphorous content estimate
U2
U3 produced by the monitoring system can be used to control
-0.9 U4
M also the nickel concentration and thus cancel this
-1
0 5 10 15
disagreement.
Process time, days
As one can see from above, the electroless nickel
Fig. 11. The online-estimated mixed potential (M) and plating reaction is a complex process in which several
equilibrium potentials of hypophosphite oxidation (1), process parameters, like concentrations, temperature,
phosphorous deposition (2), hydrogen evolution (3), and loadings, bath aging, etc., affect greatly on the reaction
nickel deposition (4) reactions. through complex, rather contradictory manners. In the
 Proceedings of the 9th WSEAS International Conference on Automatic Control, Modeling & Simulation, Istanbul, Turkey, May 27-29, 2007 64

process, a change in one parameter can easily eliminate the [2] G.G. Gawrilow, Chemical (electroless) nickel plating,
effect of several control efforts making the process (Redhill, United Kindom: Portcullis Press Ltd., 1979).
monitoring and control difficult. In this kind of process, the [3] R. Weil, J.H. Lee, I. Kim and K, Parker, Plating and
supervision of some particular bath parameters does not surface Finishing, 62 (1989).
guarantee a stable, uniform end product. Therefore, the [4] Y.A. Huang, X.T. Zeng, X.F. Hu and F.M. Liu,
supervision of some individual parameters should be Electrochemica Acta 49, 4313 (2004).
replaced by the supervision of the overall process dynamics. [5] European Parliament, Directive 2002/96/EC of WEEE,
In the case of electroless nickel plating process, this means (2003).
that instead of supervision of individual parameter, like [6] W.J. Cheong, B.L. Luan and D.W. Shoesmith, Applied
reagent concentrations, the electrochemical processes, like Surface Science 229, 282 (2004).
current densities and equilibrium potential, should be [7] G. Salvago and P.L. Cavallotti, Plating 59, 665 (1972).
supervised. If the electrochemical core of the process is in [8] Y.-S. Kim, H.-J. Sohn, J. Electrochem. Soc. 143 (2) 505
balance, also the deposited alloy is uniform. Because the (1996).
current densities cannot be controlled directly but, for [9] R. Tenno, K. Kantola and H. Koivo, Electroless nickel
example, through nickel concentration and pH-index, the plating: PCB process modelling and estimation. Proc.
individual process parameters is used for process control. In AIChE Annual Meeting (Austin, Texas, AIChE 2004).
this case the correct control action can be decided by using [10] R. Tenno, K. Kantola and A. Tenno, Electroless Nickel
process model implemented in the proposed monitoring Plating Process Model for Plating Through Hole Board
concept. Manufacturing. J. Electron. Mater, Vol. 35, No. 10,
2006, pp. 1825-1836.
[11] J.T. Winowling Jappes, B. Ramamoorthy and P.
7 Conclusion Kesavan Nair, J. Mater. Process. Tech. 169 (2) 308
In this paper a monitoring system for electroless nickel (2005).
plating process was described. The system was tested with [12] M. Goosey, Circuit World 28 (3) 36 (2002).
industrial data gathered from a PTH manufacturing line and [13] N.M. Martyak and J.E. McCaskie, Plat. Surf. Finish 83
it was shown that the complex process could be (22) 62 (1996).
characterised online in terms of the electrical, chemical and [14] S.-W. Ricky-Lee, C.C. Yan, Z. Karim and X. Huang,
product parameters, using a model-based monitoring system Assessment on the Effects of Electroless Nickel Plating
and standard measurements of the industry. Based on the on the Reliability of Solder Ball Attachment to the Bond
industrial data, the monitoring system was used for reaction Pads of PBDA Substrate. Proc. IEEE Elec-tronic
analysis of the electroless plating process. Components and Technology (Las Vegas, NV, USA
The data provided online by the system is similar to 2000).
the data acquired later from the laboratory analyses. In
addition, it enables online state estimation of the process in
terms of natural parameters and therefore provides an option
for fast failure detection and correction which is essential in
industrial processes.
The electroless nickel plating reaction is a complex
process in which several process parameters affect greatly
on the reaction through complex, rather contradictory
manners. Therefore, it was proposed, that the supervision of
some individual parameters should be replaced by the
supervision of the overall process dynamics, i.e. the
electrochemical processes. The process control is conducted
by the control of the reagent concentrations in which the
correct actions are decided by using the process model
implemented in the proposed monitoring concept.

References:
[1] G.O. Mallory, The fundamental aspects of electroless
nickel plating, Electroless plating: fundamentals and
applications, ed. Mallory G.O. and Hajdu J.B. (Orlando:
AESF Society 1990), pp. 1-56.