RP454

EFFECT OF ZINC COATINGS ON THE ENDURANCE

PROPERTIES OF STEEL
By W. H. Swanger and R. D. France

ABSTRACT
The effect of the surface alterations, resulting from the application and presence
of hot-dipped galvanized and electroplated zinc coatings, on the endurance
properties of 0.02 per cent carbon open-hearth iron and 0.45 and 0.72 per cent
carbon steels was determined by fatigue tests made with R. R. Moore rotating
beam and Haigh axial loading machines.
Rotating beam tests were made on: (a) Polished but uncoated specimens, (b)
specimens coated by the hot-dip galvanizing process, (c) zinc-plated specimens,
and (d) on specimens acid pickled as for galvanizing. Axial loading tests were
made on uncoated and galvanized specimens only.
The open-hearth iron was tested in the "as rolled" condition. The two carbon
steels were tested in the normalized and annealed condition, in the quenched
condition, and in the tempered condition, except that axial loading tests were not
made on quenched specimens.
The results of the fatigue tests are given in conventional S-N diagrams and are
summarized in a table, together with the results of tensile strength and hardness
determinations. Photomicrographs showing the structures cf the heat-treated
steels and of the two types of coatings are given.
The endurance ratios (endurance limit tensile strength) by the rotating beam
:

method of test of the uncoated specimens varied from 0.38 to 0.70; by the axial
loading method, from 0.31 to 0.59.
The decrease in fatigue limit from that of the polished uncoated materials
caused by the acid pickling was more marked in the quenched steels than in the
annealed and the tempered steels. The decrease varied from to 40 per cent.
A still greater decrease, as much as 42.5 per cent, was caused by the presence of
the hot-dipped galvanized coatings, The quenched and the tempered steels were
affected more adversely than the annealed steels.
The fatigue limits of the zinc electroplated specimens were equal to or greater
than those of the uncoated specimens.
The difference in the effects of the two types of coating is believed to be caused
by the differences in the nature of the bond between zinc and steel and differ-
ences in the structure and hardness of the two coatings.

CONTENTS
Page
I. Introduction 9
II. Materials 10
III. Testing procedure 12
IV. Results 13
V. Discussion 15
VI. Acknowledgment 24
VII. Selected bibliography 24

I. INTRODUCTION
It is generally agreed that the character of the surface of a metal is
an important factor in determining its resistance to repeated stresses.
If an endurance limit is accepted as an intrinsic property of a metal,
this limit is correctly determined only when smoothly polished speci-
mens with generous fillets are used. The damaging effects of surface
9
 . —

10 Bureau of Standards Journal of Research [Vol. 9

corrosion and of mechanically produced notches have formed the sub-

ject of numerous investigations on " fatigue of metals." The infor-
mation gained from these investigations has shown the necessity for
avoiding "notch effects" in highly stressed members subjected to
repeated stresses. Careful removal of tool marks, protection from
corrosion and use of adequate fillets at abrupt changes of section, aid
materially in realizing in practice the normal endurance strength of
metals.
Metallic coatings are frequently used on iron and steel to protect
against corrosion. It is a matter of considerable interest to know
what effect such metallic coatings may have upon the fatigue limit of
metals when damage by corrosion is not involved. From a mechan-
ical standpoint the presence of a metallic coating on a specimen of iron
or steel introduces factors which complicate this problem. There are
two surfaces, the free surface of the coating and that of the underlying
steel, the characteristics of which may influence the fatigue limit of
the composite specimen. Another factor is the endurance strength of
the coating itself. Very little is known about this property of the
various protective metallic coatings in general use, but it is probably
low in comparison with the endurance strength of steels. The nature
of the bond or interface between coating and steel is believed to have
a very important influence on the endurance properties of the com-
posite specimen. The nature of the surface of the steel, the kind of
coating and the manner in which it is applied largely determine the
character of the bond between coating and steel.
This investigation was restricted to a study of the effect of hot-
dipped galvanized and electroplated zinc coatings on the endurance
properties of low carbon open hearth iron and two carbon steels.

II. MATERIALS
Zinc coatings were chosen because they are the most commonly
used protective metallic coatings on ordinary structural grades of
iron and steel. Both hot-dipped galvanized and electroplated coat-
ings were used because of the known difference in the nature of the
bond between steel and zinc coating of these two types. Sherardized,
"galvannealed," and sprayed zinc coatings were not studied. It is
believed that the difference in the nature of the bond of hot-dipped
galvanized coatings and sherardized or " galvannealed " coatings, and
of electrodeposited coatings and sprayed zinc coatings, is one of degree
rather than of kind. It is, of course, possible that each of the above-
mentioned types of zinc coatings might affect the endurance proper-
ties of a given steel to a different degree.
The open-hearth iron and the two carbon steels were purchased
from jobbers and were not specially made for this investigation. The
chemical compositions of the three materials (ladle analyses) are
given in Table 1
Table 1. Chemical composition of steels

Phos-
Carbon Manga- phorus
Sulphur Silicon
nese

Per cent Per cent Per cent Per cent Per cent
Open-hearth iron... 0.02 0.03 0.042 0.005
0.45 per cent C steel .45 .60 .015 .040 0.18
0.72 per cent C steel .72 .31 .017 .019 .24
 B. S. Journal of Research, RP454

Figure 1. Structure of the carbon steels. Specimens were etched with nital
(alcohol containing 2 per cent nitric acid). 450 X
a, 0.45per cent carbon steel, normalized at 875° C, annealed at 800° C; b, 0.45 per cent carbon steel,
normalized at 875° C, annealed at 800° C, quenched in oil from 830° C; c, 0.45 per cent carbon steel,
normalized at 875° C, annealed at 800° C, quenched in oil from 830° C, tempered at 595° 0.; d, 0.72
per cent carbon steel, normalized at 795° C, annealed at 765° C.; e, 0.72 per cent carbon steel, nor-
malized at 795° C, annealed at 765° C, quenched in oil from 775° C; /, 0.72 per cent carbon steel,
normalized at 795° C, annealed at 765° C, quenched in oil from 775° C, tempered at 450° C.
Swangerl
France J
Endurance Limits of Zinc Coated Steels 11

The open-hearth iron test specimens were machined from the center
of 1-inch diameter hot-rolled bars and the specimens of the two carbon
steels from the center of corresponding bars of three-quarter inch
diameter.
All of the specimens were carefully machined on a lathe and finish
ground to size. They were then polished longitudinally until all
traces of circumferential tool marks were eliminated with emery
papers of successively finer grit, ending with 0000 paper.
Endurance limits of the two carbon steels, coated and uncoated,
were determined with the steels in the normalized and annealed
condition, in the oil-quenched condition, and in the tempered condi-
tion. The details of the heat treatments are given in Table 2.
To minimize any decarburization effect, an atmosphere of illuminat-
ing gas was maintained in the furnace during the heat treatments.
As a further precaution the hardened specimens were machined over-
size and a layer 0.005 inch thick was ground off the test length after
the heat treatments.
The open-hearth iron was used in the "as rolled" condition. The
microstructures of the two carbon steels in the three conditions of
heat treatment are shown in Figure 1.

Table 2. Heat treatment of carbon steels

Temperature for-

Normaliz-
Annealing * Quenching 3 Temper-
ingi ing 4

0.45 per cent carbon steel

Op O f~l F ° C °F. °C. °F. °C.
Annealed.. 1,607 875 1,472 800
Quenched 1,607 875 1,472 800 I~526 830
Tempered 1,607 875 1,472 800 1,526 830 i. 103 595
0.72 per cent carbon steel
Annealed 1,463 795 1,409 765
Quenched 1, 463 795 1,409 765 1,427 775
Tempered 1,463 795 1,409 765 1.427 775 "842 450

1 %-inch rods, heated with furnace, held 20 minutes, air cooled.

1 %-inch rods, heated with furnace, held 40 minutes, cooled with furnace.
3 Machined test bars, heated with furnace, held 20 minutes, quenched in oil.
4
Machined test bars, heated with furnace, held 60 minutes, cooled with furnace.

The galvanized coatings were applied by the research division of

the New Jersey Zinc Co. (of Pa.) by a method approximating com-
mercial practice for hot-dip galvanizing. The specimens to be gal-
vanized were first polished to the same degree as the specimens tested
in the uncoated condition. They were then dipped in a hydrochloric
acid solution (2 parts water to 1 part hydrochloric acid, specific
gravity 1.19) for two minutes and immediately into the zinc bath
held at 440° C. (824° F.). A high-grade zinc (containing 99.94 +
per cent Zn) was used. The weight of coating obtained varied from
1.6 to 2.0 oz./ft.
2
The galvanized coatings varied from 0.0017 to
0.0035 inch in thickness. This variation in thickness was probably
caused by the fact that some of the specimens had to be dipped more
than once to obtain a complete coating. The length of time in the
zinc bath, accordingly, varied from 45 to 100 seconds.^
In order to distinguish between the effect of the acid pickling and
the combined effect of pickling and galvanizing, fatigue tests were
12 Bureau of Standards Journal of Research [Voi.9

made on specimens of each material which, after final polishing, had

been dipped for two minutes into hydrochloric acid of the same
strength used for the galvanized specimens.
The specimens of the quenched 0.45 and 0.72 per cent carbon steels
which were not galvanized were dipped into a lead bath for 45 seconds
at 440° C. (824° F.) so that they had the same heat treatment as
the corresponding hot-dipped galvanized specimens. A final polish
with 0000 emery paper was given to the lead-dipped specimens before
they were dipped into the acid or tested.
The electroplated zinc coatings were applied at the Bureau of
Standards. The procedure was as follows: (a) cathode-electrolytic
1

cleaning, two minutes, at 90° C; (b) hot-water rinse; (c) cold-water

rinse; (d) pickled in sulphuric acid (2 N), two minutes at 50° C;
(e) hot-water rinse; (/) hot alkali dip without current, two minutes;

(g) scrubbed with cleaning solution, bristle brush; and (h) plated in
acid zinc bath, 24 minutes 1.5 amperes, 35° C.
The electrolytic cleaner was made up as follows: Sodium carbonate,
30 g per liter; trisodium phosphate, 30 g per liter; and sodium hydrox-
ide, 7.5 g per liter.
The
zinc anodes for the electroplating process were of the same
order of purity as the zinc used for the hot-dipped coatings. The
thickness of the electrodeposited coatings varied from 0.0021 to 0.0031
inch, which is roughly equivalent to a 2-ounce coating.

III. TESTING PROCEDURE

The endurance limit determinations were made by both the rotating
beam and the axial loading methods of stressing. The rotating beam
tests were made on R. R. Moore machines and the axial loading tests
were made on Haigh alternating stress testing machines, in which
the specimens were subjected to alternating equal tensile and com-
pressive stresses. The form of the specimens, methods of calibration
of the testing machines, and testing procedure followed have been
described previously by one of the authors. 2
Axial loading tests were not made on zinc-plated specimens as it
was believed that the effect of the electrodeposited coating on the
endurance limit determined by this method, would be of the same
order as was found for the rotating beam tests. As a further economy
in number of specimens, axial loading tests were not made on the
0.45 and 0.72 per cent carbon steels in the quenched condition be-
cause these steels are seldom used in this condition. Rotating beam
tests of electroplated specimens of the 0.72 per cent carbon steels in
the quenched condition and of the 0.45 per cent carbon steel in the
annealed and in the quenched conditions were also omitted.
Usually nine specimens were used in the determination of each
endurance limit. One specimen of each series, except the pickled
specimens, tested on the Moore machines was subjected to 25,000,000
cycles of reversed stress at the endurance limit and then restressed
at a value 5,000 lbs. /in. 2 above the endurance limit. An annealed, a
quenched, and a tempered specimen of the 0.45 per cent carbon steel
and an annealed and a tempered specimen of the 0.72 per cent carbon
1 Theplating was clone by the electrochemistry laboratory under the supervision of Dr. W. Blum.
2 R, D. France, Endurance Testing of Steel: Comparison of Results Obtained with Rotating-Beam
versus Axially^Loaded Specimens, Proc,, Am. Soc, Testing Materials, vol. 31, pt, 2, p. 176, 1931.
frame*] Endurance Limits of Zinc Coated Steels 13

steel were stripped of their galvanized coatings, after which they

were tested at a stress just under the fatigue limit determined on the
acid-pickled specimens of the corresponding materials.
The stresses applied to the acid pickled and the coated specimens
were calculated on the diameter of the polished specimen before it was
pickled or coated. The diameters were measured with a special micro-
meter capable of a precision of plus or minus 0.0001 inch. The change
in diameter caused by either the acid treatment alone or the acid
treatment and the application of the zinc coating was in all cases less
than 0.0002 inch.
The tensile strengths of the three materials were determined on
standard 0.505 inch diameter test bars, heat treated in the same way
as the endurance specimens. Hardness determinations were made on
the ends of the tensile and endurance test bars.

IV. RESULTS
The results of the fatigue limit determinations are given in Table 3,
together with the tensile strength and hardness of the steels and the
per cent change in fatigue limits caused by the pickling, by the pickling
and galvanizing, and by the electroplating. The fatigue limits are also
shown graphically in Figure 2.
Conventional 8-N diagrams for all of the fatigue limit determina-
tions are given in Figures 3 to 9.
14 Bureau of Standards Journal oj Research [Vol. 9

S . .
oo
v CC<*I *ocn

N .OQ
•S K "O
:.-«

3 ®.t3

l§.2

+ +
a
e °
W

eoo»ooooo
I I I I I

"\
ss
.S:
o oo
o
o o
oooo
o
oo
oo o
»o >o s
O
to fe

g>ooo oo«o
*>ooo6iooo'o't^
tf
I I

13 D
<3 03

Ph
JNCOOtOHXiN'
«3 .g(MCO«5-<*le<St-00

a 3 ®.t5

53i

p»S-'"
05 «

CO PQWDfQMOO
H
P3
< OM00NNON
o> O O CR C5-*f <**

Slsl rH CN <N i-l C5 P3

.26
^Tjti-Hi-ieNCNcOOO
Si*I

03 0)
3

^ T3 _J x3 ^3
*G ,J

to
2®n *n^ t-

pc®go»g
OPS41HS0)
Sw anger!
France J
Endurance Limits of Zinc Coated Steels 15

V. DISCUSSION
fatigue limits of the specimens that had been dipped in acid
The
were lower than the fatigue limits of the polished uncoated specimens
of the corresponding materials. The decrease was not uniform for
the different materials but ranged from zero for the tempered 0.45

J..

e
^ so

©If

OS 5©
"(3
©

"1
©
^^ 5o

©•
s ^
Z2E
hffi ©
©3,^->

-yir ** ^>

CS
2 © ©
g -©
39
zzn:
13 2
o £
2E
5. « © o
•<s> S.
2 y < -I K 2
u in <
Ec,
1-©
uX<
L ^
o o o o o o o o o o o o
- o 0> CO 1^ <0 (\i - P
o
Nl X)S W3d'99T 0001 -J.mil 30NVHnClN3
per cent carbon steel to 40 per cent for the quenched 0.72 per cent
carbon steel. This result was clearly a manifestation of the " notch
effect" caused by the acid treatment and was of the same nature as
the corrosion effect which has been shown by McAdam (1, 2), 3 to
have a pronounced influence on the endurance properties of metals.
8 The numbers in parentheses here and throughout the text refer to the papers listed in the selected bibli-
ography appended to this paper.

122486—32 2
 — 1 " 1

16 Bureau of Standards Journal of Research [Vol. 9

The variations in the effect of the acid treatment on the fatigue

limits are undoubtedly associated with different solubility rates of
the materials in the three conditions of heat treatment. The differ-
ence in the surface contours of the steel in different specimens of any
one series was of about the same magnitude as the difference between
the three series of specimens (acid pickled, galvanized, and electro-
plated). Figure 10 shows, in longitudinal section, typical surface
contours of the steel of the specimens after the various treatments.
The decrease in fatigue limit resulting from the acid treatment
was much greater for the quenched carbon steels than for the an-

35 + EL ECT RO PL.ATE D ROTATING

O UN COA TE D BEAM
© PIC KLE D
e ga LVA MIZ EC "?
• RE STR ES
- DIC NO T
5E
F Alt 5 «»

30
%

ifr^
>-* *.... +_ •• —
ELECTROPLATED
Z 25
-35
6
CO •^ ^
•
X
Q.
CO ^e ^' j[\. '"
UNCOATEO
rr ... _ 1

CD (p-~- *-*z—rrfr 1

-125 °N
o "X.
s
o «V
j
- 20 CKLED
I

CO
35 / VX 1/
LC A D NG
CO
id
tr O
H30
co

UNCO ATE
fL
:>r- b» "^r
25 i

j^ e*. CALVANIi"2E 3

20 4 5 7
I0 6
{

I0 I0 . I0 I0
CYCLES FOR FAILURE
Figure 3. S-N diagrams for fatigue tests of 0.02 per cent carbon open-
hearth iron

nealed or tempered steels. This is in accord with the generally

accepted idea that a hard steel with low ductility is more susceptible
to notch effects than a softer and more ductile steel.
There was a marked decrease in the fatigue limit of the galvanized
materials as determined by the rotating-beam method except for the
open-hearth iron, for which there was little if any difference (4.0 per
cent). For the carbon steels the decrease ranged from 13 to 42.5 per
cent, and was greater for the quenched and the tempered steels than
for the annealed steels. By the axial loading method of test the
decrease for the open-hearth iron was 13.5 per cent and for the carbon
Swangen
France J
Endurance Limits of Zinc Coated Steels 17

steels the decrease was much greater in the tempered steels than in
the annealed steels.
Except for the open-hearth iron and the quenched 0.72 per cent
carbon steel, the decrease in fatigue limit was greater for the gal-
vanized than for the acid-pickled material. It might be considered
that this further decrease was caused by an increased pitting or
notch effect on the surface of the steel by the action of the zinc in
the galvanizing process. That this was not the only cause is indi-
cated by the fact that galvanized specimens of the annealed and the
45 O
©
Uf -(CO/
PI CKL. ID
m D ROTATING
© G>U.VA NI2 E D BEAM
• REISTF ES SE D
- Dl D NC )T "A L
1

x S"l RIP 'EI>

40 """--
--.
.p ^ .._ UN 2OATE0
O
,-cr
£35
CO
>» fit 1

•«.. ... ©. — ej»-

tr
Ld
4 >-».
rjr PICKLED
CL

CO
30
1 — -«
CO s GAL .VANIZED
-I - ••«.—•
025 9*

§35 A
L 0/
X! A
1
L
^JG

<o
230 N
<r
H V
(0 '«. UNC'.Of kT E 3
^.9 X,
C
"V, Er _j
25 ___|
i=~ GAL.w N
l

ZE D

20 5 6 7 f
10' I0 10 I0 I0
CYCLES FOR FAILURE
Figuke 4. S-N diagrams for fatigue tests of 0.45 per cent carbon steel,
annealed

quenched 0.45 per cent carbon steels, and of the annealed 0.72 per
cent carbon steels, when stripped of the zinc coating 4 did not fail in
10,000,000 cycles in the rotating beam machines at stresses just under
the fatigue limits of the acid-pickled materials. The stripped speci-
men of the tempered 0.45 per cent carbon steel failed after 3,500,000
cycles which indicated that its fatigue limit was not much lower than
the stress at which it failed. The stripped specimen of the tempered
0.72 per cent carbon steel failed after 300,000 cycles, but at a stress
32,000 lbs. /in. 2 higher than the fatigue limit of the galvanized material.
Hence it is believed that the conclusion is warranted that the
presence of a hot-dip galvanized coating causes a serious lowering
4
The zinc coatings were dissolved in hydrochloric acid (specific gravity 1.19) containing 1 ml of antimony
chloride solution [32 g of SbCU in 1,000 ml HC1 (specific gravity 1.19)] to 100 ml of acid. A. S. T. M. specifi-
cation A 90-3Q,
 — ,

18 Bureau of Standards Journal of Research [Voi. 9

of the fatigue limit of carbon steels below that which would be ob-
tained in the same steels in the polished but uncoated condition.
A similar conclusion can be drawn from the results of investigations
by Harvey (3, 4), Haigh (5), and Fuller (6) of the protection against
corrosion fatigue afforded to steels by galvanized or other types of
metallic coatings. Although their results showed that the endur-
90 R(JT •|
NG
? III
O
9
UNCOATED
1

% BE A K1
PICKLED
» GALVANIZED
• RESTRESSED
- DID NOT FAIL
STRIPPED
85 x

\
V
\
\
o \
\
C I
s — ^-^o \
o \
80
UNCOATE D

? 75 O
6
to o
a 70
CL

<o ©
CD
-J 65
o \
o V.
^.
5

o ©
i 60
'«*,

—— -- -• --
PICKL ED
©*
to
CO
UJ
<r
a
fc
55

50 ^ \
\
\
\
\ \

'».
^
-
~9z
\ -»
45 \

^ GALV> \W. IZ

40 10' 5 7
I0 !0 6 I0 10*
CYCLES FOR FAILURE
Figure 5. S-N diagrams for fatigue tests on 0.4-5 per cent carbon steel,
quenched

ance properties of the coated metals were definitely better than for
the uncoated metals, when subjected to simultaneous stress and cor-
rosion, the corrosion-fatigue limits of the galvanized materials were
at the same time lower than the endurance limits of the uncoated
materials not subjected to corrosion.
In marked contrast to the lower fatigue limit of the galvanized
specimens, the fatigue limits of the zinc-plated specimens of the softer
Sw anger]
France J
Endurance Limits of Zinc Coated Steels 19

steelswere equal to those of the corresponding uncoated specimens.

The fatigue limits of the zinc-plated specimens of the tempered steels
were higher than those of the corresponding uncoated specimens by
5.5 per cent for the medium carbon and 11 per cent for the higher
carbon steel.

55 + ELECTROPLATE! ) ROTATING
O UNCOATED BEAM
6 PICKLED
©
©
-
GALVANIZED
RESTRESSED
DIDNOT KAIL \
^> '-
u.
» 1 I 1 1 1 1

x STRIPPED -CS ELECTROPLATED

50 ^ 'A* «».
N
~"c?~
T* i PICKLED
T*iT
XX 7rp*
UNCOATED
45 » 1 N
\
\
\
\
i
\
"Sic
240 ©~
LVANIZED
6 G/s

<n
«35
£55 L O/
AX
)l NG
L

<o
(0
-J50
o ""**"—-«„

o
o a
JN CC A T ED

'45 & 9*
(

(0
<0
Id
cc
*~4 "*L

35
4
*
•

'iisL G/^LV A Jl Z ED
30

25 ^ : C 7 {
10' J0 I0 I0 I0
CYCLES FOR FAILURE
Figure 6. S-N diagrams for fatigue tests on 0.4-5 per cent carbon steel, tempered

Both the hot-dip galvanized and the electroplated specimens had

been subjected to comparable acid treatments before the coatings
were applied. Furthermore, the steels of each series were identical
in composition and had received the same heat treatments. Hence
any differences in fatigue properties of thefgalvanized andfelectro-
plated specimens of the same series can not be ascribed to differences
in the steels themselves, particularly to any suspected differences in
 —

20 Bureau of Standards Journal of Research [Vol. 9

rate of notch propagation. It is believed that the increased fatigue

limits of the electroplated coatings were not the result of a strength-
ening effect of the zinc coating. Numerous cracks were found in
both types of coatings in the more highly stressed portions of the
specimens after they were removed from the testing machines.
This indicates that the maximum fiber stresses in the coatings were
above their endurance limits. It is reasonable to expect that the
endurance strength of the galvanized coating was higher than that
of the electrodeposited zinc. Consequently any strengthening effect

45 O
+ ELECTROPLATE.D
UNCOATED ROTATING
© PICKLED BEAM
© GALVANIZED
• RESTRESSED
- DID NOT FAIL
x STRIPPED
40 \
\. +
N» »
'
Z 35 a! .... —- —
ELECTRO >L ATED
d
<o
C£
30
UJ
£L
40
-s.
(0 > «^
UNCOATED
CO
_l 35 ©
B
>1
1 °^ l'
e
^
-.-
V
1

O \ 1

o 2f?
-
^ PICKLED
o ..."
i ii

I 30 -gr GA LV Ah IIZEO
to
Id
cc25
H 35 A XIAL
L 0/XDING

w^^,
30 .^V--
»i.'~
I JNCOAT
-
IMI
CA LV Ar JIZED

25 10' 5 6 7 (
I0 I0 I0 I0
CYCLES FOR FAILURE
Figure 7. S-N diagrams for fatigue tests on 0.72 per cent carbon steel, annealed

from the coating would be expected to be derived from the galvan-

ized rather than from the electrodeposited coating.
The two types of coating are different in many respects. The
electrodeposited coatings, as would be expected, were homogeneous
throughout. The bond between zinc and steel may have been a
"molecular bond," but at the interface there was little, if any, dif-
fusion of iron into the zinc. Figure 11 (a) shows a typical cross
section of the electroplated steel specimen.
The structure of hot-dipped galvanized coatings on steel has been
studied by a number of investigators (7, 8, 9). Without going into
Sw anger!
France J
Endurance Limits of Zinc Coated Steels 21

a detailed discussion of the composition of the layers it suffices at

this time to state that the galvanized coatings on the specimens used
in this investigation consisted of at least three layers. The outer
layer was predominantly pure zinc, the innermost layer consisted
largely of the more or less well-known iron-zinc intermetallic com-
pounds. The intermediate layer consisted largely of intermetallic
compound or compounds interspersed in a zinc matrix. On the
annealed steel and the open-hearth iron specimens the intermediate
layer was relatively much thicker than the innermost layer, whereas,
on the quenched and on the tempered steels the thickness of the
innermost layer approached that of the intermediate layer.

135 o
©
u NCO AT ED ROTATING
p CKL ED
© G ALWVNI ZE D BEAM
• R EST 3E 5S
- D D N DT FA IL

Z 130 ©
~~
6 \r~-~ — -- - — ""~~~-—
(0

Ld
125 r i .
-_--.

(L UNC OA T :c i

>*
«0
CD.
-J
120
O 85
o
o
I
80 •
(0 \
\
<0 \

Ld n K
\ \
CL N *Q G A LVA NIZED
\

H 75 — — fcit—
CO PIC KL E D

70 5 7 8
10' I0 I0 6 I0 I0
CYCLES FOR FAILURE
Figure 8. -S-N diagrams for fatigue tests on 0.72 per cent carbon steel, quenched

Micrographs (c) and (d) Figure 11 show in cross-section the gal-

vanized coatings on an annealed and on a quenched 0.45 per cent
carbon steel. The difference in thickness of the inner, iron rich,
layers is marked. Possibly a difference in solubility in zinc of the
heat-treated steel, and the annealed steel or open-hearth iron is
responsible for the difference in thickness of the alloy layers.
The scratch hardness of the two types of coating indicated that the
outer layer of the galvanized coatings was of the same order of hard-
ness as the electrodeposited coatings. The intermediate and inner-
most layers of the galvanized coatings were increasingly harder as
the. steel surface was approached. This is illustrated in the micro-
graphs (e) and (/) of Figure 11. The alloy layer adjacent to the steel
(e) appears to be even harder than the steel itself.
As stated before, there were numerous cracks in both types of
coating of the tested specimens. In the electrodeposited coatings
 —

22 Bureau of Standards Journal oj Research [Vol. 9

I 15 + EL ECT ROPL A FEC ROTATING

O UN COA TED BEAM
e pic .KLE D
e ga LVA >JIZEC
• RE STR ESSE D
- OIC NO T FAI
>
x ST RIPF
I 10
\
1 ""-
-

I 05 - "— ._...±... — JI
ELECTRO PL A" ED

I 00
\
\
\
\
\
\
\

£ 95 _o_^
V

"^2r-
6
in
o-
UNC:0ATED

- i.
£ 90 e

Q. •©.
>*.
- ^ I

jt
£0 ** CKl-EC
85 :
p '

o 60 •-
\
o \
\
o \
>
\

s \
S
»

(0
55 9 \

<0 GA LVA Nl ZE D
Id

£ 50
<0 70 .IAL
LC )A DING

65
(3
s*

ur JCC iA TED
60 >-

55
\
\
50 -_\
\
\
N;
< - ......
If* GA LVA Nl ZE.0

45 10' 7
I0 5
6 10'
I0 I0
CYCLES FOR FAILURE
Figure 9. S-A diagrams for fatigue tests on 0.72
T per cent carbon steel, tempered
 B. S. Journal of Research, RP454

*-i ''IJ '

k
iiJk*„m^*: U., fi^jm^J.. '«mHmJL*,~4k *'»„ *. ^Jj0k,^ t p,fa^Jit* * £K&k
i l .-

wgt mhmiMIm

D
f )

Figure 10. Typical contours of the steel surface of speci-

mens unetched. X 50
a, Polished; polished and then acid pickled; c, polished, acid pickled,
6,
and then galvanized; d, polished, acid pickled, galvanized, and then
stripped; and e, polished and then electroplated.
|,B. S. Journal of^Research, RP454

Figure 11
, Structure of electrodeposited zinc coating on 0.72 per cent carbon steel, annealed; b, crack in gal-
vanized coating extending into steel, 0.45 per cent carbon, annealed; c, structure of galvanized coat-
ing on 0.45 per cent carbon steel, annealed. Note the relatively thick intermediate layer, the thin
innermost layer, and the light colored outer layer of relatively pure zinc; d, galvanized coating on
0.45 per cent carbon steel tempered. Note the thickness of the inner, hard, brittle layer adjacent
to the hardened steel as compared with the thin layer of similar composition next to the softer
steel of photograph c; e, scratch made with BLerbaum microcharacter across galvanized coating on
tempered 0.45 per cent carbon steel. Note differences in width of scratch in the outer zinc layer
and the intermediate and innermost iron-zinc alloy layers. Note that the innermost alloy layer
appears to be harder than the steel; /, scratch made with Bierbaum microcharacter across elec-
trodeposited zinc coating on tempered 0. 72 per cent carbon steel
. Discontinuity of scratch between
zinc and steel was caused by difference in elevation between zinc and steel. Etched with chromic
acid solution containing sodium sulphate (20 g Cr03, 1.5 g Na2S04 in 100 ml H2O). X 275. Arrows
indicate steel base.
Swanger] Endurance Limits of Zinc Coated Steels 23

the cracks were irregular (a) Figure 11, and appeared to be inter-
granular. In the galvanized coatings there were many more cracks
in the intermediate and innermost layers than in the outer layer.
Many of the cracks appeared to have started in the layer adjacent to
the steel and to have progressed toward the outer surface, and in
some instances the surface was not quite reached. In many instances
the cracks undoubtedly originated at the surface and progressed
inwardly toward the steel.
Of the broken specimens which were sectioned for examination
under the microscope, two were found in which there was a crack in
the steel which was a continuation of a crack in the innermost layer
of the galvanized coatings. One of these is shown in (6) Figure 11.
Although there were many cracks in the galvanized coatings, the
evidence indicated that only a few had advanced into the steel. The
probability is that none of the cracks in the electrodeposited coatings
had extended into the steel, and that the cracks which caused failure
of the electroplated specimens in the fatigue test originated in the
steel itself.
The explanation of the lower fatigue limits of the galvanized
specimens as compared with the uncoated or electroplated specimens
is believed to lie: (a) In the difference in the stress conditions at the
bottom of a crack in the inner, relatively hard, layers of the galva-
nized coating and those in a similar crack in the softer electrodeposited
zinc; and (b) in the difference in the nature of the bond between zinc
and steel in the two types of coatings.
It has been shown (10) that in relatively soft and ductile metals
subjected to repeated stresses, slip lines form either previously to, or
subsequently to, the appearance of cracks and that the cracks advance
in the direction of the slip lines. The slip lines are an indication of
plastic deformation under an applied stress which decreases when the
deformation occurs. It is believed that the zinc of the electrode-
posited coatings had sufficient ductility to deform around the bottom
of an advancing crack, and that the resulting decrease in stress con-
centration when the crack had advanced to the steel was sufficient
to stop the crack at that point. The discontinuity between zinc and
steel was an additional aid in halting further advance of the crack.
Consequently, the normal endurance limit of the steel was attained.
In the case of the galvanized specimens, a crack advancing into the
relatively hard and very brittle inner layers did not meet with any
conditions conducive to a decrease of stress concentration. The
crack progressed to the outer steel fibers with undiminished stress.
The intimate bond between coating and steel offered no obstacle
to the advance of the crack into the steel. Naturally not every crack
produced in the coating in the course of the fatigue test penetrated
into the steel. A fortuitous combination of maximum stress concen-
tration and conditions at the surface of the steel most favorable to
the propagation of the stress, determined the location of the crack
which led to failure of the specimen. Consequently since the pres-
ence of a hot-dipped galvanized coating promotes stress concentra-
tions, the fatigue limit of such a coated specimen was appreciably
lower than the normal endurance limit of the steels.
The data obtained on the specimens which were restressed at 5,000
2
lbs. /in. above the fatigue limit, after they had been subjected to
25,000,000 cycles of stress at the fatigue limit, indicated that only the
24 Bureau of Standards Journal oj Research [Vol. 9

quenched and the tempered steels were appreciably strengthened by

the previous understressing. All of the annealed carbon steel and
open-hearth iron specimens tested failed to "run" at the higher stress.
The 0.72 per cent carbon steel specimen, uncoated and the two electro-
plated specimens of the 0.72 and the 0.45 per cent carbon steel, in
the tempered condition, also failed to run.
The interesting observation was made that distinct spangles were
developed on the surface of the zinc of the galvanized specimens
shortly after they were placed in operation in the rotating beam
machines.
VI. ACKNOWLEDGMENT
Grateful acknowledgment is made by the authors to P. R. Kost-
ing for his aid in the preparation of the micrographs of the zinc coat-
ings and of the scratch hardness tests.

VII. SELECTED BIBLIOGRAPHY

1. D. J. McAdam, jr., Influence of Water Composition on Stress Corrosion,
Proc, Am. Soc. Testing Materials, vol. 31, pt. 2, p. 259, 1931; and addi-
tional papers by McAdam listed at the end of this paper.
2. D. J. McAdam, jr., Stress Corrosion of Metals, Proc, Zurich Congress of the
New International Association for the Testing of Materials, 1931.
3. W. E. Harvey, Zinc as a Protective Coating Against Corrosion-Fatigue of
Steel, Metals, and Alloys, vol. 1, p. 458, April, 1930.
4. W.E. Harvey, Cadmium Plating v. Corrosion-Fatigue. Pickling i>. Corro-
sion-Fatigue, Metals and Alloys, vol. 3, p. 69, March, 1932.
5. B. P. Haigh, Chemical Action in Relation to Fatigue of Metals, Trans. Inst.
Chem. Engrs., vol. 7, p. 29, 1929.
6. T. S. Fuller, Some Aspects of Corrosion Fatigue, Trans. Inst, of Metals
Division, A. I. M. E., p. 47, 1929.
7. H. S. Rawdon, Structure of Commercial Zinc Coatings, Proc, Am. Soc.
Testing Materials, vol. 18, pt. 1, p. 216, 1921.
8. W. H. Finkeldev, The Microstructure of Zinc Coatings, Proc, Am. Soc.
Testing Materials, vol. 26, pt. 2, p. 304, 1926.
9. W. Guertler, Structure of Galvanized Iron, Zeitschrift fiir Metallographie,
vol. 1, p. 353, 1911.
10. H. F. Moore and Tibor Ver, A Study of Slip Lines, Strain Lines, and Cracks
in Metals under Repeated Stress, Bull. No. 208, Eng. Expt. Sta., Univ.
Illinois, June 3, 1930.
11. H. W. Gillett, The Need
for Information on Notch Propagation, Correlated
Abstracts, Metals and
Alloys, vol. 1, p. 114, September, 1929.
12. H. W. Gillett, What is This Thing Called Fatigue? Metals and Alloys, vol.
2, p. 71, February, 1931.

Washington, April 30, 1932.