Effect of Curing Temperature at an Early Age on the Long-

Term Strength Development of UHPC

Summary
In order to achieve compressive strengths above 200 MPa, UHPC is normally heat treated at
an early age (2 days) for 24 to 48 hours at 90°C. E ssentially, this accelerates the pozzolanic
reaction of silica fume with Ca(OH)2. Contrary to expectation, this long-term study shows that
the strength development of UHPC after heat treatment has not ceased. The compressive
strength of a heat treated UHPC was observed to increase from 220 to 280 MPa over 8
years’ storage in water. 29Si MAS NMR spectroscopy revealed that the chain length of the C-
S-H phases increased from values between 5 and 6 to 9 during this period. A distinct
increase in compressive strength of UHPC not subjected to heat treatment (20°C) was also
observed at high ages. The compressive strength was 250 MPa at an age of six years which
was 58% more than the 28 day strength of 160 MPa. This is due to the slow, but continuous,
pozzolanic reaction of the silica fume and the increase in C-S-H phase chain length.

Keywords: UHPC, silica fume, heat treatment, 29Si MAS NMR, C-S-H chain length

1 Introduction
Between 1990 and 1995 in France and Canada progress was made in the development of a
type of concrete called Poudres Réactives (BPR) or Reactive Powder Concrete (RPC). In the
mean time the term ultra high performance concrete (UHPC) has become established as the
international designation for this type of concrete. Since the end of the 1990s several UHPC
demonstration structures have been built such as filigree footbridges with wide spans in
Sherbrooke (Canada, 1997), Seoul (South Korea, 2002) and Kassel (Germany, 2007). In
Cattenom (France, 1997), a UHPC substructure of a cooling tower may be found which is
very durable despite exposure to an aggressive environment. Other examples of UHPC
realization include several aesthetically ambitious demonstration objects like the toll station in
Millau (France, 2004). These structures are all less than ten years old and therefore the
concrete still relatively young. The design of the structures was usually based on test results
for concrete specimens with an age of at most 28 days. The experience gained during the
design and realization of some of these structures influenced the French guidelines for
UHPC [i] which appeared in 2002. The results of this study on the long-term development of
compressive strength at high ages of up to eight years should increase the level of trust and
acceptance for the novel cementitious material UHPC. Since it is not possible to extract
cores from the highly stressed UHPC structural components, the development of
compressive strength at high ages is of particular interest with regard to the assessment of
safety margins.
In the present investigations, not only specimens stored under normal conditions (20°C,
enclosed in polyester foil or submerged in water) were considered, but also specimens

1
subjected to heat treatment for 24 hours at 50, 65 or 90°C at an early age. Besides
measurements of compressive strength, 29Si MAS NMR spectroscopy was used to determine
the degree of hydration of cement and silica fume as well as the chain length of the C-S-H
phases.

2 Experimental
2.1 Materials and Compositions
The present investigations were performed using four different UHPC compositions and a
model binder paste mix which were produced between 1999 and 2001. The same cement
and silica fume were used for three of the UHPC mixes stored at 20°C and the model mix
(made in 2001). The heat treated UHPC1 (produced in 1999) contained different binder
materials. The 28 day compressive strengths after storage at 20°C were between 155 and
171 MPa, see Table 1. The composition of the UHPC and the curing method are in Table 1.

Table 1: Composition and curing of the UHPC mixes

Mix UHPC1 (1999) UHPC2 (2001) UHPC3 (2001) UHPC4 (2001)

28 d compressive
*** 160 MPa 171 MPa 165 MPa 155 MPa
strength
Temperature, storage 20°C
20°C 20°C 20°C
and heat treatment 50, 65 and 90°C
CEM I 42.5 R- CEM I 42.5 R-
Cement CEM I 42.5 R CEM I 42.5 R-HS
HS HS

w/c 0.23 0.27 0.33 0.27

Silica fume content

28 30 30 18
[wt.%] w.r.t . cement

Quantity Quantity Quantity Quantity

Material kg/m³ l/m³ kg/m³ kg/m³ kg/m³ l/m³ kg/m³ l/m³

Quartz sand (< 0.5 mm) 872 329 943 356 943 356 943 356

Stone flour (< 0.063 * ** ** **

153 56 458 149 458 149 458 149
mm)

Cement 889 287 683 212 636 198 736 229

Silica fume 248 105 205 87 189 80 132 56

Superplasicizer 57 52 44 41 28 25 43 39

Water 171 171 155 155 192 192 171 171

Total 2390 1000 2488 1000 2446 1000 2483 1000

* Quartz flour
** Basalt flour
*** Stored at 20 °C
2.2 NMR Spectroscopy
Solid state NMR experiments were performed with a Bruker Avance 300 spectrometer
(magnetic field strength 7.0455 T, resonance frequency of 29Si is 59.63 MHz). To measure
the 29Si MAS NMR spectra, the samples were packed in 7 mm zirconia rotors and spun at
5 kHz at an angle of 54°44’ (MAS). The chemical shi fts were recorded relative to external
tetramethylsilane (TMS). The single pulse technique was applied with a pulse width of 6 µs.
Owing to the slow relaxation of the silica fume, a repetition time of 45 s was chosen [ii] and a
typical number of scans was 2000. Thirty Hertz line broadening was applied to all spectra
prior to deconvolution. The signal patterns of the spectra were deconvoluted with the Bruker
WINNMR software using a Lorentz curve which led to the best result. The interpretation of
the 29Si NMR spectra was performed according to the Qn-Quotation, Figure 1.

29
Figure 1: Si MAS NMR spectrum for the model binder mix Portland cement and
20 wt.% silica fume (w/c = 0.27) at an age of 28 days [iii]

By combining the integrated intensity of the different NMR signals, it is possible to calculate
the degree of hydration of cement Z and silica fume sf with respect to Si as well as the mean
chain length of the C-S-H phases, equations 1 to 3 [4].

I (Q1 ) + I (Q 2 ) + I (Q 4 ) − sf ⋅ I (Q 0 ) Eq. 1
H Si , Z , = Z
I (Q 0 ) + I (Q1 ) + I (Q 2 ) + I (Q 4 )

I (Q1 ) + I (Q 2 ) + I (Q 0 ) − sf ⋅ I (Q 4 ) Eq. 2
H Si , sf = Z
I (Q 0 ) + I (Q1 ) + I (Q 2 ) + I (Q 4 )

C =
(
2 I (Q 1 ) + I (Q 2 ) ) Eq. 3
I (Q 1 )

3
3 Experimental Results
3.1 Investigation of Hydration Using 29Si NMR Spectroscopy
The slow pozzolanic reaction of silica fume at 20°C which led to a change in the structure of
the C-S-H phases (see Figure 2) was responsible for the increase in compressive strength at
high ages. Parallel strength and 29Si MAS NMR measurements performed with hardened
binder specimens at an age of 28 days (w/c = 0.27, 20 wt.% silica fume) and UHPC at an
age of one year (w/c = 0.27, 18 wt.% silica fume) and eight years (w/c = 0.23, 28 wt.% silica
fume) confirm this relationship, cf. Table 2.
Using 29Si MAS NMR Zanni et al. [5] proved that at 20°C qua rtz flour in concrete is absolutely
inert - even after long storage times. Moreover, silica fume has a larger specific surface than
quartz flour. Thus the model mix, which is almost identical in composition to the binder paste
matrix of one of the UHPC mixes, is comparable to UHPC.

Table 2: Degree of hydration of cement and silica fume it dependence of age, heat
treatment and composition

Degree of Mean Compressive

Specimen Age
hydration chain length strength
CEM SF
Specimens stored at 20°C [days] [%] [%] [-] [MPa]
Model mix (0.27 – 20) 1 30.6 10.2 3.5 92
Model mix (0.27 – 20) 7 38.5 26.2 4.4 135
Model mix (0.27 – 20) 28 37.8 47.1 5.7 156
[years] [%] [%] [-] [MPa]
UHPC2 (0.27 – 30) 1.123 39.1 51.3 7.6 199
UHPC3 (0.33 – 30) 1.123 47.2 66.3 6.0 175
UHPC4 (0.27 – 18) 1.123 35.2 72.8 5.8 172
UHPC1 (0.23 – 28) 8.370 30.7 76.4 8.4 252
Heat treated specimens
UHPC1 (0.23 – 28) [years] [%] [%] [-] [MPa]
Temperature / Prestorage time
50°C / 1 d 8.438 25.2 85.0 7.7 263
90°C / 1 d 8.553 30.9 90.4 9.1 247
50°C / 5 d 8.427 35.4 78.8 8.3 * ≥ 256
65°C / 5 d 8.427 34.6 81.3 7.6 * ≥ 272
90°C / 5 d 8.543 35.6 83.3 8.1 249
* these compressive strengths were already determined at an age of 6.041 years
 8 years

1 year

28 days

7 days

1 day

Initial mix (0 h)

29
Figure 2: Si MAS NMR spectra for the model mix (0.27-20) up to 28 days (bottom) and
UHPC mixes after 1 year (0.27-18) and 8 years (0.23-28)

At an age of 28 days, the proportion of hydrated silica phases of the cement in the model mix
was approximately 38%. The hydration reaction of the cement was completed after, at latest,
seven days, see Figure 3. The proportion of reacted silica fume increased during normal
storage from 47%, at an age of 28 days, to 76% after eight years. During this period the
amount of C-S-H formed clearly increased and the Q2 signal for bridging silicate tetrahedra at
-85 ppm increased indicating a growth in mean chain length of the C-S-H phases from 5.7 to
8.4. In the case of the heat treated specimens, the proportion of reacted silica fume (80 to
90%) was much larger. The degrees of hydration of cement and silica fume determined by
29
Si NMR spectroscopy and the chain lengths of the C-S-H phases are shown in Table 2.

Figure 3: Hydration of the silicate phases of cement and silica fume measured with 29Si
MAS NMR for the model mix (w/c= 0.27, CEMI, 20% sf)

5
3.2 Long-Term Development of Compressive Strength at 20°C
The results for the compressive strength measurements for very different UHPC
compositions at the high ages of 1, 3.5, 6 and 8 years are shown in Figure 4. At an age of
3.5 years the strength is 40% above the 28 day value indicating a considerable increase in
hardening at high ages. Since all the UHPC mixes were produced without fibres, the
specimens were extremely brittle - especially at high ages. Because of this, specimen
preparation and mounting in the testing machine led to a large scatter of the measured
values: It was up to 50 MPa within a test series. This explains the lower compressive
strength of the eight year old compared with the six year old specimens in Figure 4; 225 as
opposed to 250 MPa, respectively.

Figure 4: Long-term compressive strength of different UHPC compositions stored at 20°C

(enclosed in a polyester sheet or submerged in water). Values are plotted versus
time on a square root scale.

The results of the 29SI MAS NMR investigations showed a distinct increase in mean chain
length of the C-S-H phases at high ages. Obviously, strength increases with the mean chain
length of the C-S-H phases. This relationship confirms the results of Zanni et al. [iv].

4 Effect of Age at Heat Treatment and Temperature on Compressive Strength

at Ages up to 28 Days
Beginning the 24 hour heat treatment at an age of five days was found to be most favourable
for the achievement of high strengths at ages up to 28 days, Figure 5. This was valid for heat
treatments at 50, 65 and 90°C, Figure 5. After five days, the maximum possible hydration of
the Portland cement is reached which is below 100% owing to the low w/c ratio. The effect of
the storage time before heat treatment is confirmed by the results of Ma [v] who
recommended commencing heat treatment on reaching a UHPC compressive strength of
90 MPa. The present UHPC compositions reached this strength in just under three days.
Depending on the type and dosage of superplasticizer, strength development can be delayed
by up to two days under certain conditions.
Figure 5: Effect of age at the beginning of heat treatment and heat treatment temperature
on the compressive strength of UHPC 1 at ages up to 28 days.

A high compressive strength of 225 MPa, obtained by optimal heat treatment (beginning at
an age of five days, 90°C) and reached after only s even days, was equalled only after a very
long storage period (roughly 3.5 years) at 20°C, cf . Figure 4.

5 Long-Term Development of Compressive Strength after Heat Treatment at

an Early Age
The results of these investigations contradict statements in the literature where the
development of shrinkage and strength as well as the chemical reactions causing them are
terminated by heat treatment at an early age. All the heat treated specimens with ages up to
eight years exhibited a further gain in strength of as much as 30%, see Figure 6.

Figure 6: Long-term evolution of compressive strength of UHPC 1 during storage under

normal conditions and after heat treatment at early ages

Irrespective of treatment history, the specimens between six and eight years old all had high
compressive strengths of 250 to 300 MPa which corresponds to values measured for natural
stone, e.g. granite.

7
If UHPC is not heat treated at an early age or only at temperatures up to 90°C then it may be
assumed that a final strength of 250 to 300 MPa will be reached after six years. Significantly
higher strengths are only possible through the more complicated process of heat treating at
temperatures around 250°C. By adding steel fibres t o the mix, typically 3 vol.%, ductility
increases and the statistical scatter of strength values is reduced. It must therefore not be
expected that fibres increase strength.

6 Conclusions
• To obtain high compressive strengths at early ages it is expedient to heat treat UHPC
for at least 24 hours following storage for five days at 20°C. After five days, the
hydration of the cement, which is 35 to 40% due to the low w/c ratio, is essentially
complete.
• Heat treatment at 50 to 65°C at an early age resul ts in high compressive strengths of
200 MPa. At 90°C the pozzolanic reaction of silica fume is greatly accelerated so that
compressive strengths of 225 MPa are reached at an age of only seven days.
• Irrespective of previous treatment, all specimens exhibited very high compressive
strengths ranging from 250 to 300 MPa at high ages of six to eight years. This
corresponds to values measured for natural stone, e.g. granite. The ongoing increase
in strength at high ages is due to the pozzolanic reaction of silica fume and growth of
C-S-H phase chains to lengths between eight and nine SiO4 tetrahedra.
• Since the design of structural components made of UHPC is based on compressive
strength measured at an age of 28 days or following heat treatment, a safety margin
of 20 to 50% is valid for concrete strength at an age of 6 years.

7 References

[1] Interim Recommendations for Ultra High Performance Fibre-Reinforced Concretes. Association
Francaise de Genie Civil (AFGC) / Service d´ètudes techniques des routes et autoroutes
(SETRA), working group on Ultra-High Performance Fibre-Reinforced Concrete, Januar 2002.
[2] H. Hilbig, F.H. Köhler, P. Schießl: Quantitative 29Si MAS NMR spectroscopy of cement and silica
fume containing paramagnetic impurities, Cem. Concr. Res. 36 (2006) 326-329
[3] H. Hilbig, F.H. Köhler, P. Schießl: Hydratation von Hochleistungs-Feinkorn-Betonen, NMR-
spektroskopische Untersuchungen, 15. Internationale Baustofftagung, Tagungsband, 2003,
Weimar, 1-0489 – 1-0496.
[4] Zanni, Hélène; Cheyrezy, Marcel; Maret, Vincent; Philippot, Samuel; Nieto, Pedro: Investigation
of Hydration and Pozzolanic Reaction in Reactive Powder Concrete (RPC) Using 29Si NMR.
Cement and Concrete Research, (1996) Vol. 26, No. 1, S. 93-100.
[5] Ma, Jianxin: Experimental Investigations for the Production of Ultra-High Strength Concrete.
Lacer No. 6, 2001, S. 215-228.