Jeffrey Thomas and Joshua Detellis,

GCP Applied Technologies, explain the

use of grinding aids in vertical roller
mills to reduce cement prehydration.

Abstract further links this prehydration to lower strengths and

Vertical roller mills (VRMs) have important advantages longer setting times. These findings demonstrate how
over ball mills for the production of cement, including reducing the prehydration of a VRM operation has the
greater energy efficiency, smaller plant footprint, and potential to significantly improve the resulting cement
the versatility to grind multiple cement types under performance.
different conditions. However, cement produced in
VRMs sometimes exhibits lower strengths and longer Introduction
initial setting times compared to ball-milled cements Since VRMs were introduced in the 1990s as an
made with the same raw materials. This degradation in alternative to ball mills for finish grinding of cement,
properties can be directly linked to prehydration of the they have gained an increasing footprint in the market.
cement, which is caused by the presence of moisture In a VRM, the particles are ground by a combination of
inside the VRM. A primary source of moisture is the compressive and shear forces as a roller passes over a
spraying of water onto the clinker as it is ground, which compacted bed of material. This is quite different from
is routinely done to stabilise the layer of clinker that a ball mill, where grinding occurs by dynamic impacts.
passes under the rollers, known as the grinding bed. As a result, operation of a VRM is more energy efficient,
Fortunately, the need to spray water can be reduced with less energy lost to heat and sound. Production of
significantly through the use of suitable grinding cement in a VRM is estimated to be about 36% more
aids that stabilise the grinding bed without causing efficient than production in a ball mill, on the basis
prehydration. This paper presents data from VRM field of tonnes of finished cement produced, which leads
trials that directly correlates VRM water spraying levels to direct cost savings. In addition, VRMs have other
to prehydration of the resulting cement and then advantages, such as a relatively small footprint and
short residence times, which allow for rapid adjustments problematic level of prehydration, versus only 19% for
to the grinding operation. ball mill-produced cement.
Despite these advantages, it is often observed by
plant operators that the strength and setting times Prehydration mechanisms and morphology
of the cement produced in a VRM is inferior to that While prehydration is a reaction between cement
produced from the same raw materials in a ball mill. In and water, it differs in important ways from normal
the past, this was often blamed on insufficient gypsum hydration. In the normal hydration process, the cement
dehydration in the VRM. However, it is now clear that, particles are surrounded by liquid water, resulting
in the majority of these cases, the cause of the problem in extensive dissolution of the cement minerals and
is prehydration of the cement, which can be broadly allowing the hydration products to grow out away
defined as an unwanted reaction between cement and from the particle surfaces. With prehydration, the
water that occurs before the final intended use of the water is generally present as a vapour that adsorbs
cement. It has long been known that prehydration has in limited quantities onto the cement surface. Under
a negative effect on cement properties and that the these conditions, only the most soluble phases tend to
magnitude of these negative effects increases as the dissolve and react. In cement, these are generally the
amount of prehydration increases.1 These negative free lime (CaO), the alkali sulfate phases, and aluminate
effects include the following: phases.
Importantly, due to the limited availability of water,
ll Reduction of compressive strengths at all ages, prehydration products tend to form as a compact layer
with the greatest reductions at early ages. around the cement particles (Figure 1). This layer tends
ll Delayed mortar and concrete setting times. to limit the further dissolution of the cement particles,
ll Reduced bulk powder handling properties, such as delaying setting and strength development. The
flowability, lump formation, and blockage. presence of the layer can also interfere with the action of
ll Reduced mortar and concrete slump. chemical additives, rendering them less effective. Thus, it
ll Inconsistent chemical additive and admixture may be difficult to mitigate the effects of prehydration
performance by using accelerators or traditional quality-improving
grinding aids. The common adjustment made by plants
In both VRMs and ball mills, water may be in response to prehydration is to grind the cement to
sprayed onto the material as it is ground, resulting in a higher fineness. This has well-known disadvantages,
prehydration of the cement that is produced. However, however, such as increased energy consumption,
VRMs tend to use considerably more water than decreased throughput, and increased water demand for
ball mills; thus cement produced in VRMs exhibits, the finished cement. In summary, the prehydration of
on average, higher prehydration levels than cement only a small fraction of the cement during finish grinding
produced in ball mills. GCP Applied Technologies can have quite significant effects on the properties of
recently conducted a survey of 181 different cements the cement when it is used to make concrete or mortar.
produced in both ball mills and VRMs.2 The survey Mitigating these effects after the prehydration has
found that 48% of the VRM-produced cements had a occurred can be difficult.

Figure 1.Scanning electron microscopy (SEM) images of two cement particles. Left: cement with minor prehydration
– the cement surface is mostly smooth and free of hydration products. Right: cement with severe prehydration – the
surface is completely covered with a dense layer of hydration products that will act as a barrier, reducing the reactivity
of the cement when mixed with water.

Reprinted from October 2018

World Cement
Causes and mitigation of VRM prehydration be noted that the material on the grinding bed consists
There are a number of factors that make VRMs more not only of the relatively coarse raw material that has
susceptible to causing cement prehydration compared just entered the mill, but also partially ground material
to ball mills. Probably the most important of these that is much finer, but has not yet passed through
is the practice of spraying water onto the grinding the separator. Factors that affect the quality of the
bed of a VRM in order to stabilise the bed and reduce grinding bed include the material quantity (and bed
mechanical vibrations, which can cause damage to the height), particle size distribution, bulk mass flow, and
equipment. This continuous introduction of moisture compressibility. Of particular importance is the amount
into the mill is certainly a major factor in prehydrating of very fine material in the bed. The presence of fines
cement. Other factors include recirculation of moist has a negative effect on grinding: it is believed that the
process gasses, which is done to keep the VRM outlet fine particles distribute the compressive forces in the
temperature up and limit the use of external heating, grinding bed in such a way that less particle fracturing
and lower operating temperatures, which can limit occurs, making the grinding process less efficient.
evaporation and increase the relative humidity As noted above, water is typically sprayed onto
compared to a ball mill. the VRM bed. This water causes the finest particles to
Another issue that is related to prehydration is silo agglomerate, which effectively makes them behave
set, which can occur when cement undergoes gypsum more like large particles in terms of the behaviour of
dehydration followed by additional prehydration the bed. This practice is quite effective at reducing
after being stored in the silo. In a VRM, low operating mill vibrations and allowing for stable VRM operation;
temperatures, high humidity levels, and short residence however, spraying water onto the bed causes cement
times often result in minimal gypsum dehydration. prehydration. The more water sprayed the higher the
Therefore, VRM outlet temperatures should generally degree of prehydration.
be maintained at a low enough level such that the A more effective way to stabilise the grinding bed
cement is not above about 70°C when it enters the silo. is by the use of a process grinding aid applied either
A good recent discussion of VRMs and silo set is given directly to the grinding bed or to the raw feed as it
by Canut and Theisen.3 enters the VRM. The use of a grinding aid to stabilise
There are two broad strategies for reducing the bed allows the water spray to be reduced, and
prehydration in a VRM. The first strategy is to reduce may also allow the production rate to be increased.5
the amount of water sprayed onto the VRM table Grinding aids are believed to stabilise the bed by
through the use of a grinding aid that stabilises the dispersing the fine particles so that they do not stick to
grinding bed. This is discussed in more detail below. the larger particles or to each other. As a result, they
The second strategy is to perform various operational pass more readily through the separator and into the
adjustments to the mill. While VRM operation is cement product, so fewer fines are returned to the
quite complex and there are a number of factors and grinding bed. As a result, the bed is more stable and the
strategies that can be effective, the primary goal is grinding process is more efficient.
to increase the amount of moisture that is removed In general, grinding aids for ball mills tend to also
from the mill in the exhaust gas stream. This typically provide benefits for VRMs.5 In particular, the dispersing
entails increasing the amount of fresh, relatively dry air effect tends to benefit both types of systems. However,
brought into the mill, while reducing the recirculation the operation of a VRM is quite different from that
of moisture-laden process gasses.4 of a ball mill in many respects, so it is not surprising
At some plants that manufacture blended cements, that grinding aids formulated for ball mills are not
a significant amount of water may be introduced into optimal for VRMs. GCP Applied Technologies has used
the VRM in the raw feed. Natural pozzolans appear to its extensive experience with both types of grinding
have the greatest capacity for bringing in moisture, but systems to develop a new product line of grinding aids
limestone and even gypsum may also have appreciable that are specifically formulated for VRMs, known as
moisture. In terms of prehydration, the total water TAVERO™ VM grinding aids.
introduced by spraying and in the raw feed is what
matters. Experimental methods
In general, both strategies (use of a grinding aid
and process adjustments) can be implemented for best Measuring prehydration
results. Their benefits are demonstrated by the results Prehydration is measured by heating a cement sample
of two field trials discussed later in this paper. and measuring the weight loss within a defined
temperature range. As originally defined by Thiesen
VRM grinding aids to reduce vibrations and Johansen, prehydration was quantified as the
The smooth operation of a VRM is highly dependent parameter Wk, which is the percentage mass loss of
on the characteristics of the grinding bed, which is a sample as it is heated to just before the portlandite
defined as the layer of material on top of the grinding starts to decompose (about 375°C), but adjusted
table. The VRM rollers pass over the grinding bed, to remove the contribution to the mass loss from
compressing it and causing particle fracture. It should gypsum and plaster dehydration.1 Prehydration is

Reprinted from October 2018

World Cement
 most accurately measured using a thermogravimetric
analysis (TGA) instrument. Figure 2 shows TGA data
for three VRM-produced cements with varying
degrees of water spray, and therefore varying levels
of prehydration. Shown are both the derivative curve
(rate of mass loss vs. temperature) and the cumulative
curve (mass vs. temperature). It should be noted that
other temperature ranges could be used to measure
prehydration; for example, the small weight loss before
gypsum dehydration may be excluded.2 Naturally,
the resulting values will be different, but the trends
obtained should be similar.

Sample collection and testing

Results reported here were measured using samples
Figure 2. Thermogravimetric analysis of three cements collected during field trials conducted at two VRM
produced in a VRM with different amounts of water plants. After adjusting the operating parameters of
spray. The peaks in the differential weight loss curves the VRM, such as feed rate, additive dosage, and water
represent decomposition of specific phases: GYP-gypsum spray rate, to the desired levels, the VRM was run for
and plaster, SYN-syngenite, and CH-portlandite. Wk at least 30 min. to ensure stable operation and that
is the percent mass loss up to the start of portlandite the cement being produced represented the current
decomposition, excluding the gypsum dehydration. From conditions. Approximately 3 kg of cement was collected
the cumulative curves, it can be seen that Wk increases from the output belt and, once cooled to near room
as the sprayed water increases. The Wk value for the temperature, the cement was vacuum-sealed in plastic
3.5% water sample is the sum of the two regions at left bags using a small vacuum sealer. The samples remained
labelled “Wk”. sealed until the time of measurement. All reported

(a) (b)

(c) (d)

Figure 3. Results from VRM field trial 1, illustrating the improvements in cement quality that can be obtained by
reducing prehydration (Wk). a) Prehydration of cement samples, plotted as a function of the water spray level.
b) Initial setting time, plotted as a function of prehydration level. Mortar compressive strengths at two-days (c) and at
28 days (d) plotted as a function of prehydration level.

Reprinted from October 2018

World Cement
(a) (b)

(c) (d)

Figure 4. Results from VRM field trial 2, illustrating the improvements in cement quality that can be obtained by
reducing prehydration (Wk). a) Prehydration of cement samples, plotted as a function of the water spray level. b)
Initial setting time, plotted as a function of prehydration level. Mortar compressive strengths at 1-day (c) and at 28-
days (d), plotted as a function of prehydration level.

compressive strength and setting time results were Figure 3, this decrease in prehydration resulted in
measured following EN 196-1 and EN 196-3 testing strength increases of approximately 3.5 MPa at both
standards, respectively. 2 days and 28 days, as well as a decrease in initial
setting time of about 45 min. These are significant
Results performance improvements, given the relatively
modest change in prehydration.
Field trial 1
The first case study is a field trial at a VRM plant Field trial 2
manufacturing a Type I cement, consisting of 90% The second case study is a field trial conducted at a
clinker, 5% limestone, and 5% gypsum. The moisture plant manufacturing a blended cement containing
content in the raw feed was minimal. Cement 74% clinker, 22% natural pozzolan, and 4% gypsum.
production rate was 176 tph at the start of the trial, Production rate was 87 tph. In this case, there was a
but increased to just over 200 tph by the end of the high raw moisture content of the pozzolan of 10%,
trial. Water spray onto the grinding table was 4.4 % resulting in about 2 tph of water being introduced
by weight of raw feed at the start of the trial with the into the VRM from the raw feed. Water spray onto
use of the plant’s traditional ball mill grinding aid. The the grinding table was 3.6 tph at the start of the trial.
mill outlet temperature was maintained at 85°C. The total initial water input was thus about 6.4% by
After replacing the plant’s current grinding mass of feed, which is almost 50% higher than for
aid with a TAVERO™ VM grinding aid, the water Case 1. The mill outlet temperature was 87°C.
spray was progressively reduced to 2.6%, while After starting the application of a TAVERO™ VM
also making other process adjustments to the mill. grinding aid, the water spray onto the table was
Cement samples were collected and sealed for later reduced from 3.7 to 2.1%. Figure 4 shows the
analysis during this process. Figure 3(a) shows the relationships between water spray, prehydration,
prehydration of the cement as a function of the total and cement properties. Notably the measured Wk
water input (expressed as a percentage by mass of values are roughly three times the values measured
the raw feed). It was possible in this trial to decrease in Case 1. This is mostly due to the water content and
the prehydration level (Wk) by 0.08%, when water Wk contribution of the pozzolan. Despite the high
spray was reduced from 4.4 to 2.6%. As shown in level of prehydration, the benefits of reducing the

Reprinted from October 2018

World Cement
prehydration by 0.04% helped improve set time by References
20 min. and 1 and 28 day strength by 1.7 and 2.3 MPa, 1. THEISEN, K. and JOHANSEN, V., “Prehydration and
respectively. Interestingly, the cement with the Strength Development of Portland Cement”, American
lowest water spray had the second highest Wk value Ceramic Society Bulletin, 54(9), (1975), pp. 787 – 791.
but, as seen in the figure, the setting time and the 2. SILVA, D., THOMAS, J., KAZMIERCZAK, D., and CHEUNG,
compressive strength of the cement samples follow J., “Prehydration of Cement: Global Survey and
the measured Wk values. This indicates that water Laboratory Results”, ZKG International, 71(6), (2018), pp.
spray is not the only factor for prehydrating the 55 – 60.
cement. If the pozzolan being added to the mill has a 3. CANUT, M. and THEISEN, K., “Cement Grinding:
range of moisture contents, this could cause a cement Silo‑Safe Cement”, World Cement, (January 2017), pp.
sample taken with low water spray to still exhibit a 49 – 52.
high level of prehydration. 4. MARSAY, K., GIBSON, L., and CHEUNG, J. “VRM
Optimization”, ICR, (August 2017), pp. 56 – 60.
Conclusions 5. MARSAY, K., “Grinding Additives for VRMs: Myth or
Cement produced from VRMs are much more likely Reality?”, World Cement, (April 2014), pp. 139 – 143.
to suffer from prehydration compared to cement About the authors
produced from conventional ball mills. This is due Jeffrey Thomas is a Senior Principal Scientist at GCP
primarily to the additional water spray used to stabilise Applied Technologies in Cambridge, Massachusetts, US.
the VRM grinding bed. A prehydrated cement will have Previously, he was a Principal Scientist with Schlumberger
deteriorated properties, such as lower compressive and a Research Professor at Northwestern University. He
strength and longer setting time. In order to reduce has authored more than 70 scientific articles in the area of
water spray in a VRM, grinding aids can be used to cement chemistry.
help stabilise the grinding bed. In the two field trials
discussed here, replacement of a standard grinding Joshua Detellis is a Chemist at GCP Applied Technologies
aid with a grinding aid optimised for VRM grinding in Cambridge, Massachusetts, US. He received his master’s
have been demonstrated to allow for a substantial degree in Chemical Engineering from the University of
reduction in water spray, with corresponding decreases Massachusetts. He has 7 years of experience working in the
in prehydration level. As a result, compressive strengths cement industry.
were increased by up to 5 MPa and initial setting time
was decreased up to 45 min 

Reprinted from October 2018

World Cement