O
O

344

 

THE EFFECT OF AIR ENTRAINMENT
UPON COMPRESSIVE STRENGTH
OF CONCRETE

11min For H10 Dunc d B. S.
MICHIGAN STATE COLLEGE

Kent A. Allmeior
1948

1 'I 1'11-

.g. ‘9...

 

v. a J.» ik 191's: 2‘. t'o 0‘ H

n. p. 3...... [JIIP .5 E.”

..§ ’I.‘l. I mol‘o- .rSI-V
tkg..§%¥t

5.7 5&th

. II.
3|Jlo ll.- .. E...
I...

. I. :U.‘f’ 3... Id.- u‘mv (3?» V! Vgt‘lnlta'. £515.31“? m-

IDJJI i,

I’l‘. . .‘Q
. t

\

_

.... 2 a.

The Effect of Air Entrainment Upon

Compressive Strength of Concrete

A Thesis Submitted to

The Faculty of
MICHIGAN STATE COLLEGE
of

AGRICULQURE AND AFFLIED SCIENCE

by

Kent A.‘£;1emeier
Candidate for the Degree of

Bachelor of Science

June 1948

e/n‘w

Of”

Acknowledgement

 

I shall take this Opportunity to extend sincere thanks
to Mr. M. V. Esrdiello, Director of the Toledo Testing Lab-
oratory, whose suggestions caused me to write upon this
subject, and to Mr. J. T. FcCall of the Civil Engineering
Department for his cooperation and assistance throughout
the past two months in the writing of this paper.

Appreciation must also be extended to the Research
Division of the Michigan State Hifihway Department and to
the Metallurgical Department for their fine cooreration
and also for the use of their equipment.

Preface

The process of purposefully introducing air into
concrete came primarily as a remedy for surface scaling
of concrete pavements due to freezing and thawing and to
arrlication of chlorides. Exhaustive tests have rroved
that it has accomrlished its purrose quite admirably.

And the benefits of this fifth ingredient, air, do not
stor with merely imrroved durability for highway rave-
ments. Entrained air also improves workability of the
fresh concrete, reduces segregation, and hence reduces
bleeding or water gain. But all this benevolence has not
been rotten free. It must be paid for and the rrice is

a reduction of the comrressive, tensile, flexural, and
bond strengths of the hardened concrete. Small additions
of air over and above ortimum quantities cause serious
reductions in strength and the absolute control of quant-
ities of entrained air is still in a developrent stage.

In pavement work discrepancies of this nature have not
proved prohibitive. But flexural strength is not affect-
ed as severely as is the comnressive strength. the quest-
ion would then arise as to whether air entraining agents
may be used as freely in structural work as they are being
used in highway work without stringent and exacting con-
trols over quantities of air entrained. It is the under-
standing of the author that over the east two or three

years the use of air entraining concrete has come to be

known as a cure all for all types of concrete ills. This,
of course, is not true. With this in mind I have exrlored
a small phase of the disadvantages of air in concrete.

I shall present the material in two rarts. The first

part presents a result of library research of the broad
subject; the second part rerorts the findings of labor-
story work concerning reduction of strengths caused by

entrained air.

Background

 

Although the benefits caused by abnormal quantities
of air voids in concrete have been recogniaed off and
on for a number of years it has been only during the last
decade that air entrainment has proved its real worth.
In the late 1920's a Professor C. A. Scholar, of Kansas
State College, noted that cement containing a metallic
stearate produced concrete having greatly imrroved re-
sistance to cycles of freezing and thawing. He also ob-
Served that this concrete of surerior durability possessed
a subnormal unit weight. During the early 1930's tests
were rerformed in which air was deliberately entrained in
concrete. In every case resistance to freezing and thaw-
ing was found to be improved -- arrarently in direct rro-
portion to the reduction of unit weight. Alttough this
new imrroved durability was much to be desired it seemed
inserarable from the horrible nrosrect of lower unit weight.
For up until, and during, this time the best concrete was
the most dense concrete. Because of this deep rooted
density concert the idea of intentionally lowering unit
weight was out of the question. In the years following,
however, the surface disintegration of concrete ravements
arrarently became so serious that the then dormant lower
unit weight theory was again revived. However, even at

this time (1935-38) it was not generally accented that
-1-

The entrained air was the cause for imrroved durability.
Nany attributed it to some unkosn action of the admixture.
Between 1938 and 1940 several test roads were built as a
part of a growing country-aide research program. Since
1940 a great deal of conclusive investigation has been
carried on. Air entrainment in concrete is now hailed by
some as the greatest discovery since the water-cement ratio
law. fiuch has been learned about the subject -- much is
yet wanting to be known.

Any question as to whether imrroved durability found
in the less dense concrete is due to air or merely to the
agent has been answered. Tests run by H.L.Fennedy made
a comparison between three concrete sixes. One mix was a
rlain concrete containing the normal amount of arrroximately
one percent air. The other two were mixes emrloying an
air entraining agent which caused an above-normal percentage
of air. From one of the latter mixes the entire air con-
tent was removed by vacuum -- even most of the normal one
percent. Durability tests of the srecimens showed: the
mix with additional air was greatly imnroved; the mix with
an identical amount of agent but with the air removed show-
ed little or no imrrovement. In fact, the mix from which
more than the added air was removed showed decreased dur-
ability. The effect of entrained air seems to be inderen~

dent of the means used to gain the added air.

How does eir produce an improved concrete? What is
its function? The theories answering these questions seem
to be in cloee agreement and quite widely accepted. The
millions of eIenly diepereed minute air bubblee immobilize
the mixing water. This causes reduced capillary and water
channel etructure ehich in turn increases durability by
vainimizing migretion or water into and out of the hardened
concrete. Ruinoue percolating chloridee thich are applied
to pavements for ice removal are hindered. H.L.Kennedy
edvencee the pictorial ex- ¢flkued dhrflmc

L‘ "on r f
ple‘netion ehown in f1g.1 0f / / L c e e

 

 

 

     

  

how the tiny eir pockete re-

duce eurrece disruption due

‘.

%
p6;

"U
[<2

ff’aggl o’fmm2 th‘jCJ éffejpf

to freezing. The presence

of eveileble eir voide re-

 

lievee the inwerd preeeure

 

ee freezing progreeeee.

 

 

Thie preeeure tende to low-

 

 

 

 

 

 

Bwr-
er the freezing point of the

enter. If eufficient voide Concert: fl/r‘ ma [Judy/”an?
are provided for the water faykl

to creep inward ee freezing progreeeee. dieintegration of
the structure is prevented. In addition to enhancing con»
crete durability. entrained eir also improves workability
end pleceability of freeh concrete. Honeycombing end eeg-

regation ere thereby partially eliminated. The multitudes
~3-

of tiny bubbles act almost as ball bearings -- they lub-
ricate the mix. The majority of these air spheroids have
a sieve size ranging from the 50 to the 100 sieve. It is
Obvious that each bubble is comrarable to a grain of fine
aggregate. Such is exactly the case. The bubbles con-
stitute an additional aggregate in the mix possessing com-
plete flexability of share. Because the air imparts these
qualities to the fresh concrete 3 reduction in both sand
and water is possible. In fact, it is advisable to make
this reduction when incorroratina air in a normal mix in
order to maintain the cement factor per cubic yard. A re-
duction in water also regains a part of the lost strength.
Let us look briefly to the agents which are used to
entrain air in concrete. Virtually all of them depend
upon a foaming action for the entrainment of air. Vinsol
Resin, Darex. Avr-trap. and Orvus are a few of the many
commercial agents being offered today. To date, Vinsol
Resin and Darex are the only ones which have been accepted
and incorporated in A.S;l.¥. specifications. Among those
used in early studies were natural wood resins, animal and
vegetable fats. various wetting agents, water soluble soars
of resin acids, etc. It seems safe to say that Vinsol Resin
is the most popular agent in use at the rresent time. ,
with the admittance now that the rrocess of purposeful
entrainment of air in concrete has been far from completely

harnessed, let us examine the variable factors. Why is it

-4-

that so much difficulty is encountered in successfully
predicting nir content obtained in field work? Certainly
it is in the field that close control is most to be desir-
ed. Stanton *alker and D.L. Bloom have compiled a list of
these variables which seems to incorporate and agree quite
well with other avHilable writings on the subject. The
following was presented in an article in the ACI Journal,
V0117, do 6, June,1946:

"It is generall? conceeded that the following factors
are responsible for major variations in air content when
the percentage of air entraining agent is held constant.

1.) Richness of mix ~- lean mixes entrain more air than
rich mixes. Percentage of air has been found in some
cases to vary from a to 85 when using an air entraining
cement containing a given percentage of air entraining
agent, cement content varying from four to seven sacks
per cubic yard.

2.) Percentage of sand -- combined aggregates contain-
ing high percentages of sand will entrain more air then
the same total aggregates containing lower percentages

of sand. In one case on record. increasing the sand per-
centage from 38 to 42$ by weight of total aggregate in-
creased the entrained air from 4.0 to 5.5%.

5.) Sand gradation -- a deficiency in ~50 to +100 mesh
sand, or conversely an excess of +30 or «100 mesh sand

will‘reduce the rercentage of entrained air by as much

as 31.7 g ‘

4.) Consistency -- high slump concretes (up to Gin.)
will entrain more air than low slump concretes. all
other conditions being the same. Increasing the slump
from 2" to 4" will increase the entrained air aprrox-
imately let. The same slump increase in high slump con-
crete, say six-inch. will not affect the percentage of
entrained air as much.

5.) Temperature -- temrerature of the concrete as con-
trolled by atmosnheric temrerature and temperature of

the mixing water. aggregates. etc.. has a slighteffect
on percentage of air entrained. The higher the temper-

-5-

ature, the lower the percentage of air. A range of
anrroximately 2% may be exrected between temperatures
of 40°E and GO‘F.
6.) Length of mixing and type of mixer -- large or
small variations in air content may result from these
variables, derending upon the tyre of mixer and type
of air entraining agent used.
7.) Age of cement -- where air entrainment is accomr-
lished bv the use of air entraining cement, age of the
cement has been found to have a small effect on the
amount of air entrained in the concrete, depending upon
the tyre of air entraining agent used.
There is at least one more variable which perhars should
be added to this list. Charles E. Tuerrel notes in the
June, 1945 issue of the Crushed Stone Journal that crushed
stone has been observed to entrain more air, using identi-
cal mixes, than does gravel. He illustrates a case in
which the gravel concrete entrained 4.4% air contrasted
to a 5.43 entrainment with the stone.

Special note should be given to items 2 and 3, con-
cerning the amount and gradation of sand in the mix. The
sand phase of the air entraininfi mix commands a high priority
in all recent discussion. Kennedy states that the amount
of air entrained in concrete by a fixed rercentoge of air
entraining agent is a function of, among other things, the
percentage of sand by volume of the concrete. In a serarate
bulletin he contends that for a given percentage of air
entraining agent, and constant mixing conditions, the

quantity of air entrained by the efigregate is a function

of the particle size, the crtimum range being between the

No. 50 and No. 100 sieve sizes. Tests seem to indicate
that either excessively course or excessively fine sands,
as orrosed to well graded sands, will decrease the effect-
iveness of the air entraining agent. Proper control of
the sand phase is an imrortant part of the design of an
air entraining mix.

Another of the more interesting variables is the
richness of the mix. Tests have shown that a neat cement
raste incorporating an air ontrsining agent generates
but a small rercenta36 of air even with very thorough mixc
ing. The resultant paste, instead of being more blastic,
is actually less rlastic. The paste assumes a stiffer
consistency. It could seem from this that the cement in
the mix hinders, rather than side, air entrainment. This
theory is proven in the actual concrete mixture. The richer
mixes, those with a higher cement factor, consistently en-
train less percentages of air than do the leaner mixes.
These facts can only reflect back as proof of the previous
Paragrarh in which it was said that the entrained air is
a function of the aggregate. This conclusion is axiomatic
since everything else has been eliminated.

Length of mixing time is a much discussed variable
affecting an air entraining concrete. This problem would,
in all probability. have been of little concern had it not
been for the ever growing transit-mix industry. Until

recently the condition of extended mixing causing excessive

-7-

amounts of entrained air was rarticularly bothersome in
the case of Vinsol Resin. Previously, Vinsol Resin was
interfiround with the clinker or added at the mixer in its
rlpin condition. To effect air entrainment, the plain
Vinsol Resin had to react with alkalies, formed when the
cement and water combined, in order to become available
as a foaming agent. In other words, the Vinsol Resin be-
came sodium resinate at the time of mixing. Obviously,
the comrleteness of this chemical reaction was a function
of the mixing time. ihis condition has now been elimin-
ated by and large now by neutralizing the Vinsol Resin
(making into a soap - the sodium resinate) before it is
added either at the cement factory or at the mixer. how-
ever, even neutralized Vinsol Resin, called NVX, still
shows some inconsistencies due to mixing time when it has
been interground with the clinker. lhe variation seems
to be derendent upon the individual cement. vhen NVX is
added at the mixer the above mentioned inconsistencv is
greatly reduced. ihe effect of extended mixing time with
Darex seems to closely approximate conditions found with NVX.
With this background of the subject in mind let us
turn to a laboratory consideration of the possible results
which may come from any of the several discussed variables

affecting air content in concrete.

~8-

Laboratory Work

 

The primary purpose of this raper is to show how air
entrainment in concrete affects the strength of the con-
crete. You have seen the relative unnredicableness of air
entrainment; let us now examine the effect that these
variations might have uron the resulting concrete. The
task was undertaken with vengeance, with no apologies to
any quarter. In the normal design of an air entraining
concrete mix care is taken to rartially comrensate for a
reduction in strength caused by the additional air. A
standard state highway mix, for example, is redesigned.
The percentage of aggregate, usually only the sand, is
reduced to counteract the bulking effect of the air and
thereby maintain the original cement factor. The increas-
ed workability due to the lubricating action of the air
rermits a reduction of the water-cement ratio, This,too,
regains part of the lost strength. ihese factors are all
considered by the trained concrete technician or concrete
engineer. But certainly evervone who is using air en-
traihing cements and air entraining agents today is not
aware of all these facts. A bag of air entraining cement
is Just as common now as was a bag of plain cement ten
years ago. Because of its improved qualities rertaining
to workability, segregation, and water gain, air entrain-

concrete has come to be used in every tyre of work.

For this reason it is interesting to see iust what
the results are when a plain concrete mix is turned into
an air entraining concrete merely by the addition of an
air entrnining agent. This is “hat hes been dhhe in the
laboratory. A 4500 psi. plain concrete mix was designed
to which was added various rercentsges of neutralized
Vinsol Resin (NVX).

The aggregate anplysis is given in tables 1, 2, and
3. The course aggregate was a very smooth, well rounded
gravel. The mix used was l:2.4:3.5 by weight with a water
cement ratio of 5% gel/sack. Tyre I Huron rortlsnd cement
was used. The NVX which was in the powdered form res wade
into solution so that 1000c of solution contained one gram
of EVX. It should be recognized that this is not a true
one rercent Vinsol Resin solution since the NVX hes a Vin-
sol content of arrroximately 88%. The remainder is rude
up of the 6% caustic neutrnlizer rlus moisture pnd non-
volatile oils. The first batch was the control batch which
contained no added sir entreining agent. Six 6"x12" test
cylinders were made from each batch ~~ three to be broken
at 7 days and three to be broken at 28 days. After setting
24 hours at room conditions the cylinders were placed to
cure in a standard moist closet maintained by the'richignn
State Highway Derartment. ‘

A mixing time of 6 minutes was used. It was learned

from the Highway Department that at least six minutes was

-10-

Aggregate Anagxsis

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Fine Course
s
Percent
Free roisture 2.?fi * 1.22% *
Bulk Specific
Gravity 2.67 2.69
Unit Weight
per cu. ft. 112.5 lb. 109.5 lb.
*values varied
Sieve Analysis
Fine Course
Percent Percent Percent Percent
Siege Retained Passing Sieve Retained Passing
# 4 0.62 99.38 1%" 0 100
# 8 10.81 89.19 1” 10.13 89.87
#16 29.94 70.06 5/!" 25.60 74.40
#50 50.91 49.09 1/2" 51.00 49.00
#50 80.67 19.55 5/8" 76.75 24.25
#100 97.31 2.69 No.4 100 k 0

 

 

 

 

 

 

 

 

~11-

 

required for thorough mixing in this rarticular machine.
An undue amount of difficulty was exrerienced with the
mixer. The machine originally had been of the double oren
end, horizontal tyne powered by a gas engine. It has now
been converted to a single oren end tilting tyre with
electric motor. The orening, however, is so swell that a
funnel has been devised with which to charge it. In using
one half of the normal six cylinder batch as a buttoring
batch, the machine so robbed the mix of mortar that the
discharge consisted almost solely of course aggregate.

ihe barrel of the machine is so intricate and clogged with
hardened concrete that emrtying is very difficult. ihis
condition has come about because the machine discharges so
incompletely and sluggishly. If there is to be one do-
finite conclusion drawn from this work, that conclusion is
the definite need for a remedy of the situation.

Aside from the mixer the only other difficulty en-
countered was with the varying surface moisture of the
aggregate which was gotten from outside stockriles. It was
interesting to note that a small increase in the surface
moisture, which increased the W/C ratio and produced a wet-
ter mix, caused a marked reduction in the resulting air
content. ihis fact seems to be diametrically crposed to
the general rule (see item 4, rage 5). Because cf this,
strict control had to be maintained over the aggregate

from day to day.

~12-

The air content of the fresh concrete was calculated
by means of the theoretical weight method. Following is a

samrle comrutation:

'?ercent Air 3 Theoretical Wt.(air free) - Wt. Concrete
Theoretical wt.?eir free)

 

The theoretical air free unit weight of the concrete
is found by dividing the total weirht of the connonent
ingredients in the batch by the total absolute volume
of the comronent ingredients in the batch, ie,

total WHtht
of components = cement 29.25 lbs
FA 70.25
CA 102.50
water 14.00

216.00 lbs

total absolute

 

 

 

 

volune comronents = cement ?9°95 8 .1499 cu ft

3.13 x 82.4

FA 70-95 g .4320
2.67 x 62.4

CA _ 109.? : .6120
2.69 x b2.4

water 1400 : .2942
l x 62.1

1.4070 cu ft

30, theoretical air free
unit weight = .I3%§7_.; 153.5 lbs/cu ft

iherefore, rercentqge of
air in concrete : 153.3 - “sighed Unit ”eight X 100
- 153.3

3 band“.

238 a} soc-d.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

#- 3 min .masdwm a
9365.30 n no amuse»:
«m9 ohm .u mind in” omwm. 3:2
, omen 33 L. o 69 a .5 83. 24.4
mums mm" L. 9mm...” dm 82. 1...:
«won a." .m. 9d: m.~ owns. an:
meme meme L: mid m.m 88. mo:
momwl men .3 m9: 9: H8. 1...?
flan 3mm .3 0.03 We owoo. so:
owe: «8.... .w Toma mA 38. n44
3:. 3R 2 9%” n.” 58 ~12
sea can. i. as an a. than”. .3...
.namuanm assoc rfiwfikm «88 3. so}: a E «

 

 

-14-

Table 4 gives a complete tabulation of the percentage
of NVK used in each batch, the percentage by volume of air
entrained, the slumps, unit weights, and resulting strengths
at 7 and 28 days. The curves in Fig. 3 show the strength
relations at 7 and at 28 days as the air content is increas-
ed. Figure 3 presents a graphical ricture of the marked

reductions in compressive strength due to the entrained air.

-15-

N .bk
.920\.NV.V.NNK \‘szmtkhao Q\T
2&on \fl \ttustx c ukmxbteb {buxtx \o \ttxkou xxs

 

 

N\ \\ %\ m Q R V M: V M) N \ Q

q . fl 4 w . q a . . e + Q w
a”
J
a
m
w
1 Q%\ a
r
Ir
4 J
w
.6
Ir
1 seem y
la
4 r

l QQ§N

 

J 89“.

 

 

A
f

toq-

— 7 04]:
’° ’ - ----- 26 DayJ

80*-

rcducr/ax m Comp. J/rcnjffl

7.?
r
“u—
Jb—-
E
’ w r°"1
i‘ ' -
40" "
K

 

 

 

Q‘ 70 /.0 ’4 J3“ 4-5 7‘ ‘45. 70 28 % Id] '4 ll./ 'o
poxenfaye: 0/ um I» addl’fion *0 +6.: )mrma/ AI);

IWEVTCVF/VFP IHEVSMAIIWNJAAS [AV \57rfi£7~%?779'
r9. .7

-17-

Conclusion

 

Figure 3 is conclusive in itself. Strength reductions
of such a high order are unallowable even in pavements, to
say nothing of structural work. It should be borne in mind
that the mix used in these tests was a comraratively rich
mix. A leaner mix would have entrained greater quantities
of air and reduced the strength more.

Mr. E. Hayfield, Sales Supervisor for the Hercules
Powder Company, makers of Vinsol Resin, wrote the follow-
ing in his letter of April 27, 1948: "So offset the in-
creased bulk and to maintain the same cement content per
cubic yard of concrete, it is customary to reduce the fine
sand in the mix. It is our understanding that concrete
mixes so designed frequently have strengths only about 5
percent less than would be obtained with normal portland
cement.“ Mr. Hayfield areaks of an air entraining cement
which is manufactured to entrain an optimum air content of
from 3 to 5 percent. A reduction of only 5 percent in
strength is negligible -- such a small reduction is absorb-
ed by factors of safety. Under ideal conditions the re-
duction may be held down to 5 percent. In fact, in very
lean mixes it has been shown that the allowable slash in
water content due to the added air has actually caused a
higher strength. But we have seen how the air content may

inadvertently vary several percentage points, and we have

-18-

seen how great an effect an addition of one or two rercent
of air can have upon the comrressive strength. Of course,
if the variance acts to lessen the amount of entrained air,
‘the consequence is not as imrortant.

Until the science of air entrainment in concrete be-
comes more exacting, probably the best solution to the
problem is the discreet use of air in any work where strength,
not durability, is the more imnortant factor. ihe various
interested parties who are conducting tests on the subject
are new strongly promoting a step that will bring air entrain-
ment closer to more exacting control. That step is the use
of air entraining agents added at the mixer in favor of the
use of air entraining cements which have had the agent inter-
ground with the clinker. Control of the air has been proved
to be more accurate by this method.

Perhaps one of the biggest hazards in the use of air
in concrete is the temptation that it offers for economy.

An excellent arrearing concrete can be made with a very low
cement content if sufficient added air is entrained with
the mix. But strengths would be so low that despite the
relief of stresses provided.by the added air, the concrete
would not be serviceable. Strength is still a requirement
of portland cement concrete, and air is not a substitute

for portland cement.

-19.

Yhotomi crogrerhs

 

0n the following pages is presented a series of photo-
micrograph: showing the structure of the hardened concrete
specimens made in this test. It is interesting to note
the step by step increase in porosity. The first six micro-
graphs are enlarged 8 times. The last two. which compare
the normal concrete with the same mix with 12.4% air, are

enlarged 14 times.

Photomicrogrephe Showing

Structure of Hardened Concrete‘

"

 

Specimen AT-Z
No NVX added
1.3% air
ground surface
magnified 8x

AT-4

.oosfi uvx

2.3% eir
ground surface
8x

AT—6

.Olfi NVX

5.97- 311'
ground surface
Bx

 

AT-B

7.8 air
ground surface
8x

AT-lo

.03% nvx
11.4% air
broken surface
Bx

AT-ll-R

.05% NVX
12.4% air
broken surface
8x

 

 

 

.22-

AT-Z
No NVX added

1.3% air
broken surface
141
« 3“”a;;¢-'
A“ ‘3’? ‘13.
AT—ll-R
.05% uvx
12.4% air
broken surface
14x

A striking comparieon.between a normal
concrete and a concrete which contains
12.4% entrained air. Note the extreme

spongy texture of the lower micrograph.

.23.

' I“ .

 

 

 

 

Mm 335% cm .'

’_..-.'- —

MICHLISAH SMTE UNIVERSITY LiBRARIE—S

~‘ I HIII IIgIII
II III I ll IIIIIHI m
31293 0837

03056