ANNUAL TRANSACTIONS OF THE NORDIC RHEOLOGY SOCIETY, VOL.

27, 2019

Rheological properties of cellulose ethers and their application in cementitious

tile adhesives formulation

Fabio Curto, Matteo Monaco, and Stefano Carrà

Mapei S.p.A., Milan, Italy

ABSTRACT INTRODUCTION
Cellulose ethers (CE) are very important Tile adhesives are formulated
components of cementitious mortar cementitious mortars that have the duty of
formulations, in particular of cement-based assuring good and durable adhesion between
tile adhesives, since they determine and tiles, of whatever sort they might be, – wall
control their wet properties such as tile, ceramic tile, porcelain tile, etc. – and
workability, open time and sag resistance. various types of substrates, concrete
Cellulose ethers are either used “pure” in plywood, water-proof membranes, etc.
formulations, or modified, i.e. mixed with Adhesion is a property of the mortar once
“rheology modifiers”, such as starch ethers, – by effect of the cement hydration reaction
polyacrylamide, inorganic modifiers – it has hardened. However in the tilers’
(bentonites, sepiolites) that essentially have evaluation of the product its “wet” properties
the role of controlling the low shear viscosity are obviously even more important because –
of the product. all together – they determine its ease of
Actually, all wet properties depend on the application.
rheological characteristics of the product’s Typical “wet” properties of a tile
wet mix. The main feature of cellulose ethers adhesive are workability, open time and sag
themselves for what concerns cementitious resistance. They are all, generally speaking,
mortar formulation is a rheological attribute, rheological properties. The role of
i.e. the viscosity of a 2% solution in water, controlling them is delegated by the use in
that depends on the molecular weight of the formulations of cellulose ethers. The most
cellulose ether. commonly types of celluloses ethers in tile
In this paper several rheological adhesives formulations are hydroxy-ethyl
characterizations will be discussed, varying and hydroxy-propyl. They are either used
parameters like the cellulose ether viscosity pure or “modified” by a mix of starch ether
and its modification with starch ether and (SE), polyacrylamide (PAA), inorganic
polyacrylamide through a simple thickeners (bentonites, sepiolites, etc.) that
experimental design. The correlation among have the role of increasing the low shear
rheological properties of cellulose ethers and viscosity of the wet mix.
their behavior in a standard formulation of a The main feature of pure cellulose ethers
cementitious tile adhesive will be then themselves is a rheological property, i.e. the
studied using multivariate analysis tools viscosity of a 2% solution in water.
(Principal Component Analysis). In this paper, initially, we studied the
viscosity effect of Methyl Hydroxyethyl
Celluloses (MHEC) in aqueous solutions1.

53
F. Curto et al.

Moreover, the rheological profiles detected Table 2. Tile adhesive formulation

on these systems has been compared to those Component Dosage (w/w)
measured in alkaline solutions prepared at pH CEM I 52.5R 30.00 %
≃13 in order to simulate the typical Silica sand 67.95 %
environment of a cement paste. Re-dispersible
After these preliminary tests, the effects 1.00 %
polymer
of MHEC were investigated in cementitious Calcium formate 0.60 %
tile adhesives formulations and, together with Rheological modifiers 0.45 %
physical-mechanical properties like specific
gravity, sag resistance and wetting capability, Silica sand used has a 0.1-0.4 mm
we measured their rheological properties via granulometry, which is the most typical used
flow curves. in tile adhesives.
Finally, all the results obtained on Redispersible polymer powders are
different formulations of cement-based polymers emulsion that have been converted
adhesive with rheological measurements and by spray-drying into powdered thermoplastic
conventional mechanical tests have been resin materials that, when mixed with water,
used to find correlations using a multivariate can re-disperse back into emulsions with
analysis tool such as the “Principal essentially identical properties to the original
Component Analysis”2. copolymer emulsions.
The most common polymer compositions
MATERIALS AND METHODS used for tile adhesives are vinyl-acetate
copolymers, with ethylene or vinyl-
Materials verstatate, with a "# typically between 0 and
Three different cellulose ethers, one 30°C, depending on the copolymer
starch ether and one polyacrylamide have composition. Redispersible polymers
been investigated, identifying them by their influence hardened properties such as
declared nominal viscosity in a 2% solution: adhesion and mortar deformability but they
also decrease the viscosity (mainly at low
Table 1. Rheology modifiers shear) of the wet mix, but in this paper all the
Water retainer Viscosity (mPa·s) formulations contained the same quantity and
CE-A 3000 type.
CE-B 15000 Consistently with several previous works
CE-C 25000 carried out on cementitious tile adhesive
SE 100 systems3,4, the composition of the rheology
PAA 10000 modifiers mixtures has been determined
using the Experimental Design approach and
These rheology modifiers agents have choosing a “Central Composite Design”5, for
been hence employed in a standard a total of 18 experiments, summarized in the
cementitious tile adhesive formulation, Table 3. The use of an Experimental Design
prepared with pure clinker Type I cement, approach has been chosen because it allows
according to the following recipe: the generation of an efficient modelling of
the whole investigated system, using a
chemometric software developed at Mapei
laboratories and available with free license6.

54
ANNUAL TRANSACTIONS OF THE NORDIC RHEOLOGY SOCIETY, VOL. 27, 2019

Table 3. Central Composite Design of retainers’ mixtures

1 2 3 4 5 6 7 8 9
% PAA/Rh. Mod. 0 0 0 0 6.6 6.6 6.6 6.6 3.3
% SE/Rh. Mod. 0 0 33.3 33.3 0 0 33.3 33.3 16.6
CE Viscosity 3000 25000 3000 25000 3000 25000 3000 25000 15000

10 11 12 13 14 15 16 17 18
% PAA/retainers 3.3 3.3 0 3.3 3.3 3.3 3.3 0 6.6
% SE/retainers 16.6 16.6 0 16.6 16.6 0 33.3 16.6 16.6
CE Viscosity 15000 15000 15000 3000 25000 15000 15000 15000 15000

CE Solutions preparation calibrated pycnometer, proceeding then with

The CE aqueous solutions were prepared the EN 120048 testing of the cementitious tile
at 2, 1, 0.5 and 0.25% of cellulose by mass of adhesive.
water, dissolved with mechanical agitation The slip resistance of the tile adhesive has
for 3 min in a blender. Then, a slake time of been evaluated applying two tiles of different
24h in laboratory conditions (23°C) has been size and weight (100 mm x 100 mm at 200 g
imposed before any measurements, in order and 150 mm x 150 mm at 650 g) over a thin
to obtain a homogeneous solution free of any layer of product spread on concrete slab with
entrained air bubbles. Furthermore, the a 6 mm x 6 mm square notched trowel and
solutions with a 2% concentration were also then keeping the substrate in vertical position
prepared using a saturated calcium hydroxide for 20 minutes. The so-called “adjustability
solution in order to study the influence of pH. time” has been evaluated too, by applying
porous tiles over the cementitious tile
Adhesive mortars preparation adhesive and controlling the amount of time
The cementitious tile adhesive necessary to stick the tile in a non-reversible
formulations are prepared as a mechanical way.
mixture of all their powdered ingredients, Wetting capability was evaluated by
homogenized during a 20 minutes mixing. applying the mortar on a concrete slab and
Hence, these powders have been mixed with putting on top of it standard porous tiles
the required amount of water according to (defined once again by EN 12004) after 5, 20
EN 120046 procedures: the adhesive is mixed and 30 minutes using 20 N weights for a
for 30 seconds in a planetary mixer and, after standard time of 30 seconds and then
a manual cleaning of its bowl, it is re-mixed overturning the correspondent tile and
for another 60 seconds. The sample then visually evaluating the percentage of
observes a 10 minutes slake-time before its adhesive transferred from the substrate to the
final 15 seconds remix. tile itself.
Finally, two different operators have been
Cementitious tile adhesives test methods asked to express their judgement about the
A Brookfied viscometer has been used in workability of each formulation, evaluated
order to define the correct amount of water by troweling the products on concrete slabs,
necessary to obtain the desired consistence with a mark from 1 (worst) to 10 (best).
of the mortar (≃ 500.000 mPa·s). After this
initial check, the specific weight of the
mixture has been obtained by filling a

55
F. Curto et al.

RHEOLOGY PROTOCOLS correlated to the slip resistance test described

Rheological measurements were carried out in the previous paragraph as it simulates the
using a strain-controlled rheometer (Anton ability of the mortar to keep a tile vertically
Paar Mod. 302) equipped with different tools, still despite its weight8.
depending on the system texture which is
being considered. The CE solutions were RHEOLOGY RESULTS
made by a cone-plate geometry (diam. 50 At first, the flow curve tests were performed
mm, truncation gap 0.1 mm); the tile on solutions with 2% of MHEC by mass of
adhesive cementitious based formulas were water to verify the swelling capacity.
measured through the “ball measuring Moreover, the solutions were also prepared
system” (ball diam. 15 mm) and plate-plate with pH 12, using a calcium hydroxide
with both rough faces (diam. 50mm and saturated solution, in order to simulate the
1,5mm gap). All these tools are dedicated to cellulose behavior in presence of the
ensuring the border’s conditions for a correct cement9. The superimposition of the curves
rheological characterization. At the means, the MHECs are not affected by the
beginning the measurements were focused on alkaline solution (Fig. 2).
the “swelling” effect in water with the flow
curves measurements at 23.0 ± 0.2°C in a
shear rate range from 0,01 to 1000 s-1 in 120
s.

Figure 2. Solutions at pH 7 and 12

Figure 1. Ball system measurements For all tile-adhesives, a basic-formulation of

dry mix mortar was employed, where a mix
of rheology modifiers was added. The mix
The device for the cement-based sample “ball was composed of polyacrylamide, starch and
system” consists of a large cup where the MHEC. The water demand of the mix was
mortar has been placed and the tool with a assessed with a preliminary test performed by
ball in a non-coaxial position applies the a viscometer: the water amount is the one that
shear needed to measure the flow curve (Fig. allows the mortar to have a ‘Brookfield’
1). This device has been designed to avoid viscosity around 500 Pa·s, which
problems that occur when composite corresponds to which tilers commonly apply
materials are characterized, in particular, it these products.
guarantees the “non-slip” condition and Flow curves were performed on the sample 1,
avoids the problem related to a compressed 2 and 12 (Fig. 3), and they are all very close
material by ensuring a normal force close to to each other, due to the fact that the mixing
zero during all measurements. Furthermore, water was changed in order to obtain the 500
a creep test was performed on adhesives Pa·s Brookfied viscosity: sample 1 made
applying a constant stress of 300 Pa for 3 with the MHEC A, requested 21.5% of
minutes. This measurement protocol is mixing water, sample 12 with cellulose B

56
ANNUAL TRANSACTIONS OF THE NORDIC RHEOLOGY SOCIETY, VOL. 27, 2019

was mixed with 23% of water and sample 2

with cellulose C needed the 26%.

Figure 4. Mortars with blended CE

Figure 3. Mortars with pure CE

The creep test was performed on the same
formulations (1, 3 5 and 7) by applying a
All show a pseudoplastic profile which is constant stress in order to verify that the
typical of cement-based systems: a solid up deformation measured is related to the yield.
to a critical value called yield, corresponding By doing this it was found that there are
to low shear viscosity10, afterwards as the proportional inverses between the
shear increases the material is subjected to an deformation and the yield so sample 7 has a
internal structural rearrangement and the creep value around 10% and sample 1 is
flow occurs. Viscosity profiles in general, 1000%.
are very comparable in the medium high
shear-rate range; at 1 s-1 the viscosity is found
between 400 and 650 Pa·s.
This shear corresponds to the shear at which
Brookfield measures viscosity: flow curve
viscosities fall very close to each other at this
range, because we varied mixing water in
order to target the same Brookfield viscosity.
However, rheological tests through a stress-
controlled rheometer allows to have a more
detailed description of the rheological Figure 5. Creep: mortars with blended CE
behavior of the mix.

The real difference between samples 1, 3 5 All composition samples which are showed
and 7 is represented by the low shear in Tab. 1 were also characterized by
viscosity. This value can be tuned by rheometer: the behavior found was
changing the composition of the rheology comparable to samples 1, 3, 5 and 7 and the
modifiers: samples 3 and 5, increased it since results are summarized in Tab. 4.
the 6.6% of polyacrylamide and the 33.3% of
starch was added respectively. The adhesive
sample 7, formulated with 6.6% of
polyacrylamide and contemporarily 33.3% of
starch shows the higher low shear viscosity
value: 35·104 Pa·s.

57
 F. Curto et al.

Table 4. Experimental results

EN Heavy Op. Op.

Mix Spec. Wett. Wett. Visc
Test Adj. Tile Tile 1 2 Def. LSV Plat.
water weight 20' 30' 1s-1
Slip Slip Eval. Eval.
- % g/cm³ min mm mm - - % % % (Pa·s) (Pa·s) (Pa·s)
1 21.5 1.42 60 5 10 6 8 50 30 3580 5936 3.8 3.64
2 26.0 1.4 70 5 10 10 9 65 50 4200 3855 108.1 54.90
3 22.0 1.56 18 0.3 0.6 6 7 60 20 22.3 22010 2.5 1.82
4 24.0 1.45 25 0.6 1.8 9 8 70 35 57.5 13160 37.1 12.74
5 26.5 1.46 12 0.3 0.6 5 7 80 50 43.6 15740 9.6 7.06
6 28.5 1.36 40 0.8 4.5 10 9 90 50 91 17740 118.3 48.97
7 26.5 1.56 8 0.3 0.6 6 6 70 30 10 35390 7.6 4.38
8 26.5 1.45 15 0.3 0.4 10 10 85 30 29.6 20730 47.8 17.08
9 27 1.4 30 0.9 2.5 8 8 90 80 260.6 14210 18.8 10.57
10 26.5 1.42 30 0.7 1.5 9 9 90 75 31.6 14110 18.8 11.18
11 27 1.41 30 0.6 1.9 8 8 90 75 64.2 11730 19.2 11.70
12 23 1.39 60 5 10 6 7 75 60 3810 4455 22.6 16.31
13 26 1.49 18 0.3 0.7 6 7 75 40 49.25 16710 4.7 3.58
14 28.5 1.43 25 1.1 2.5 10 10 95 80 77 11760 79.1 34.40
15 26.5 1.4 50 1.9 9.6 9 9 70 40 1230 12300 20.6 13.92
16 25.5 1.46 18 0.6 1.8 7 7 85 75 31.5 16360 14.4 6.88
17 23 1.45 45 1.5 4.1 9 8 80 40 63.6 12410 13.0 8.87
18 28.5 1.43 18 0.5 1 9 8 95 90 43 14860 25.6 13.31

PRINCIPAL COMPONENT ANALYSIS this plot and showing totally independent

AND DATA MODELLING behaviors.
Principal Component Analysis2 is the The lower-left side of the graphs shows the
most used multivariate data analysis area with pure CE in the formulations
technique and the software11 can be resumed (experiments 1, 2, 12) and this is clearly
by two charts, the so-called “Loadings plot” highlighted by the presence of adjustability
and the so-called “Score plot”, the first and slip tests that show their higher values in
showing the correlations among the different presence of non-modified CE. The presence
tests taken in account and the second one in this area of the creep test results (Deform.)
showing the relative positioning of each indicates the excellent correlation among this
single experiment in this multivariate space. test and the mechanical behavior in EN
The comparison between “Loadings plot” standardized tests.
and “Score plot” is the key to determinate the On the right side of the graphs, in opposition
influence of each parameter on the to the tests described in the previous
investigated responses. paragraph, Low Shear Viscosity can be
Generally, directly correlated responses found, in correspondence of the retainer
are close in the “Loadings plot” while mixtures with the highest modification and
inversely correlated tests are diametrically with the lowest base viscosity, showing an
opposed. On the other hand, uncorrelated opposite behavior to those of the unmodified
results are orthogonal among themselves in CE.

58
 ANNUAL TRANSACTIONS OF THE NORDIC RHEOLOGY SOCIETY, VOL. 27, 2019

Figure 7. Loadings plot Figure 8. Scores plot

Orthogonal to these two groups a third cluster water mix ratio with a low air mounting in the
of experiments (wetting capabilities, product could be the responsible of its
workability according to the operators, water unpleasant appearance and workability.
mix ratio and CE solution plateau) can be All these results are confirmed by the model
found. The corresponding experiments are responses obtained from the mathematical
the ones with a high base viscosity and with elaboration of the whole Design of
high amounts of polyacrylamide, indicating Experiment. In particular, the behavior of slip
that a higher amount of water in the and creep test is clearly equivalent and
formulation improves the workability opposite to that of the adhesive’s low shear
properties of the mortar and its capability to viscosity.
keep a good wetting capability during time.
All these properties are contemporarily CONCLUSIONS
opposite to the specific weight of the paste, In this work the effects of MHEC was
confirming, as already known, that a low investigated on the workability and

Figure 6: Slip, creep deformation and yield in function of SE and PAA % with a CE viscosity of
15000

59
F. Curto et al.

performances of cementitious tile adhesives. 4. M. Monaco, S. Carrà, 2019, Cement

The choice of an Experimental Design that quality influence in cementitious tile
considered the dosage of MHEC and their adhesives formulation, Drymix Mortar
viscosity as input variables, allowed us to Yearbook 2019, drymix.info 4-17
reach optimal sag resistant formulations 5. R. Leardi, 2009, Experimental design
avoiding a high number of empirical tests in chemistry: A tutorial, Analytica Chimica
and, consequently, time and money. Acta 652 161–172
6. M. Monaco, R. Leardi, 2019,
Tests focused on the viscosity enhancing BasiCAT, freely downloadable from website
effect in aqueous solution showed that, http://gruppochemiometria.it/index.php/soft
within the investigated range, their behavior ware
is not affected by the pH typically found in a 7. EN 12004-2:2017, Adhesives for
cement paste ceramic tiles - Part 2: Test methods, CEN
Moreover, physical phenomena that norm
influence rheological behavior of a 8. T. G. Mezger, The rheology
cementitious adhesive has been studied and a handbook, 3rd Revised Edition, Vincentz
clear inverse correlation between zero-shear 9. C. Brumaud, H. Bessaies-Bey at all,
viscosity and creep-deformation has been Cellulose ethers and water retention, Cement
found as well as a direct correlation between and Concrete Research 53 (2013) 176-184,
creep-deformation and tile slip resistance. Elsevier
10. C. Brumaud, R. Baumann at all,
ACKNOWLEDGMENTS Cellulose ethers and yield stress of cement
We thank Michela Deola form Mapei pastes, Cement and Concrete Research 55
Rheology laboratory for her support and her (2014) 14-21, Elsevier
tips during the drafting of this paper. 11. R. Leardi, C. Melzi, G. Polotti, CAT
Special thanks to all the technicians of (Chemometric Agile Tool), freely
Mapei cementitious tile adhesives laboratory downloadable from website
for their cooperation and their evaluations of http://gruppochemiometria.it/index.php/soft
products workability. ware

REFERENCES
1. D. Bülichen, J. Kainz, J. Plank,
Working mechanism of hydroxyethyl
cellulose (MHEC) as water retention agent,
Cement and Concrete Research 42(2012)
953-959, Elsevier
2. R. Leardi, 2000, Methods of
Treatments of Data, Handbook of water
analysis – Third Edition, CRC Press, 47-76
3. M. Monaco, S. Carrà, 2018,
Chemometrics application in cementitious
tile adhesives formulation, Drymix Mortar
Yearbook 2018, drymix.info 42-60

60