JOURNAL OF RESEARCH of the National Bureau of Standards- Co Engineering and Instrumentation

Vol. 69C, No. I, January- March 1965

Soil Resistivity as Related to Underground Corrosion

and Cathodic Protection
w.

J. Schwerdtfeger

(August 4, 1964)
Corrosion data disc ussed ill this paper arc based upon m eas ure ments mad e on abo ut
4500 specimens of commonly used plain wrou ght ferrous mat ri al wh ich had been buried
in back-filled Lre nches at 86 National Bureau of Standards sites for periods up to 17 years.
The soils ran ge d in resistivi ty from 50 to 54, 000 !1-cm a nd in p H f rom 2.6 to 10.2.
Maximum pit depths at 5 years of ex pos ure a re ta ke n from pi t depth- t ime cur ves and
the curves arc also extrapolated t o probable pit depths at 30 years for eac h of the 86 sites.
Furthermore, data O il t he s pecime ns are a dju st ed to m ax imum pi t depths that might be
expected on a lar ger area, equivale nt to t hat of the exterior surface of a 20 ft length of 8
in . ull coated wrought ferrou s pipe. Maximum pe netration ratcs a nd pit depths arc presen ted
with r es pect to soil resistivit.v a nd pH.
\Veight losses which rcs ulted from about 2 years of unde rg round ex posure a rc converted
to corrosion current densities a lld after this period of expos ure current densit ies arc calculated
from the slopes of weight loss-time curves for eac h site. On t he basis of r atios of protective
current t o corrosion curre nt obtai ned from polari zation curves on steel s pec imen s underground a nd in soils in the labo ratory, t he corrosio n current densities can be a dju sted to
approximate current densit ies necessary for cathod ic protection .

1. Introduction
Between the years 1922 and 1952 the Nationftl
Bureau of Standftrds exposed tbousttnds of metallic
specimens at numerous underground sites throughout the United States. Corrosion rates of O\Ter
300 vftrieties of protected and unprotected metals
and alloys were evaluated about every 2 year s for
exposw-e periods as long as 17 years. The accumulated corrosion data were compiled by Romanoff
[IV The engineering significance of early NBS
data pertaining to some co mmonly used wTought
ferrous materials from the original 47 test sites WftS
discussed in considerable detail by Logan [2]. The
present paper is similar but includes data from about
40 additional sites and, based on the total available
data, offers s uggestions 011 the current densit ies
probably required for cathodic protection.
In this paper, in addition to extrapolating the
average maximum pit depths to 30 years as Logan
[2] did, the weight losses of the specimens are converted to corrosion current densities. From these
data the approximate current densities required for
the cathodic protection of bare undergro und surfaces
can be estimated. Although, based on NBS data,
there is no precise relationship between the corro si \Tity and resistivity of soils, a general relationship is now revealed by the total ftccumulated data.
This is also indicated by the data of Scott [3].
Recently, Schaschl and .Marsh [4] discussed t he
effect of resisti \rity on the corrosion rates of steel
in aqueous environments and showed t h at for
resistivities above 400 Q-cm, the corrosion rate bears
an inverse relation to the resistivity of the electrolyte. Thus, a study is undertaken of the influence
of soil resistivity, a meaSLll'ement readily made in the
field. The co mbined relation of pH and soil resistivI

Figures in brackets indicate the literature refe rences at the end of this paper.

71

ity to corrosion rate was also investigated. The

metftls to which our data apply are confined to
open-hearth iron , h and-puddled and mechanically
puddled wrought irons, open-hearth and Bessemer
steels, all witho ut added alloying constituents.

2. Corrosiveness of Soils
The corrosiveness of a soil might be defined as its
destructi \Te or deteriorating effect on a metallic
surface, ftS mea ured by weight loss and pit depth.
A soil Cftll be potentially cOl'l'osi \'e and yet have
a negligible effect on phin ferrous mftterials. This
was rece ntly shown by Romanoff [5] as a result of
his examination of steel pilings (pulled) exposed
from 7 to 40 years in a variety of soils some of which
hadresisti \'ities as low as 300 Q-cm. No appreciable
amount of corrosion was obselTed on pile surfaces
adjacent to soils into which the piles had been
dri\Ten below the water table, referred to as "undisturbed" soil. This type of exposLll'e is in contrast
to the soil exposure of metftls in back-filled trenches,
referred to as "disturbed" soil. On driven piling,
t he corrosion rate decreases rapidly as the initially.
available oxygen is depleted by the corrosion process
and the accessibili ty of more is limited. In distmbed soils oxygen is more readily replenished and
the soil is only very slowly restored to its natural
state. This gives rise to differential aeration and
then the corrosiveness of the soil depends largely on
its ph ysicftl and chemical properties. Properties
of the soils to be considered have already been
described [1].
The most severe corrosion usually takes place
at low elevations in poorly drained disturbed soils,
such as clays and tidal marshes, where, although
also poorly aerated, the differential-aeration effects
are large because of soil shrinkage. Then too, salts
accumulate in these areas and increase the conduc-

tivity of th.e soil. Th~ pH of such soils is usually

neutral or I? the. alkalme range. In contrast, the
least corrO SIve sO lIs are well aerated well drained
an~ seem to be in areas of high annu~l precipitation
~hlCh causes the salts to be washed away and
mcreases the soil resistivity. The pH is usually
between 4 and 7.

the .average maximum pit depth for a unit of time.

Agam , the useful form of the relation becomes ,

3. Effect of Area and Duration of Exposure

The data used in this paper are mo stly for accumulated exposure periods of 12 and 14 years. In
some soils the accumulated periods are 7 and 17 years.
The exposed area of each specimen was about 0.4 ft2.
It is of importance to know the effect of larcrer
areas and longer periods of exposure. These v~ri
abIes have been considered by previous investigators
[6,7]. Years ago, Scott [6] observed that pit depth
is a function of the area of metal exposed. Based
on National Bureau of Standards data and on data
from pipelines, a linear relation was obtained when
the logarithm of the average of the deepest pits
(se,:"eral specimen ~) on a given area was plotted
agamst the logarIthm of the area for increasing
areas. Thus, for a modest extrapolation of the area
Scott suggested the empirical equation,
'
log P = a log A+log b, or P = bAu,

(1)

where P is the average maximum pit depth on the

~roposed area A, ~xponent a is the slope of the best
Imear pl.ot , and b IS the average maximum pit depth
for a umt area. The useful form of this equation is,

P Z/P l = (A 2 /A I )a or PZ = P I (A 2/A 1)U,

(2)

where PI is the averag~ maximum pit depth on a

known area AI, and P 2 IS the extrapolated or calculated average maximum pit depth on some assumed
area Az for the len gth of exposure time. Scott suggested as a standard area that of the exterior of a
20 ft length of 8 in. diam pipe which is about 45 .16
ft 2.. The v.alues for. the exponent, a, applicable to
Spe?lmenS m 47 . sOlIs have been given by Logan ,
Ewm g, and DenISon [7]. Based on their data the
avera~e value. of a= 0.1 5 has been suggested for
use ~th all ~oils, as the exponent is apparently not
assocIated WIth any known soil property [81.
The effect of length of exposure time on rates of
corrosion underground is better known than is t h e
effect. of area exposed. For example, it is well known
that III some soils, after a few years, corrosion seems
to ?ease. The effect of duration of exposure on corrOSIOn rates depends on chemical and physical propert~es of the s?ils. Here again , it was observed that
a lmear relatIOn ~ppears to apply reasonably well
b~tween the logarIthm of the average of maximum
PIt. d~pths and the logarithm of the exposure period .
ThIS IS expressed by the equation,
l og P = n log T + log k, or p = kTn,

(3)

~here

P is. the aver age maximum pi t depth at some

tIme T , n IS the slope of the line of best fit , and k is
72

where PI i~ the average maximum pit depth for the

shor~er pen?d of exposure T I , an d P 2 is the average
ma~nmum pIt dep.th at some longer time T z. Logan ,
~wlllg, and Demson r7] divided 47 NBS soil sites
mto 4 groups based on aeration and on drainacre
r~nging from well-aerated sands and sandy loams\~
tIdal m~rshes . The exponent, n, was derived for
each soil group and r anged from n = 0.19 for soils
having goo~ aeration to n=0.68 for soils with very
poor aeratIOn, such as clays and tidal m arshes.
More recently, Scott [9] has proposed that the average !'ate of corrosion underground is inversely proportIOnal to the square root of the duration of
exposure time, that is, in eq (3), n = 0. 5.

4. Discussion of the Data

4 .1. Maximum Pit Depths
. The. maximum pit depth data for 86 soil sites are
gIve n III table 1 and graphically presented with respecFt .to soil resistivity an.d pH in figures 1 through
5. .' lgure 1 shows maXImum pit depths on the
speCImens at 5 years taken from pit depth-time curves
(not sh.own). On the average, pits are somewhat
deeper III the soils with resistivities below 500 n-cm
than in soils with higher resistivities but above 500
n-cm th.erc appears to be no regular variation bet wee!l pIt. depth an~ soil resistivity. Howe ver , the
relatIOnshIp looks dIfferent when the rate of ma,,1.mu.m. ~enetration a~ter 5 years is plotted versus
reSIstIvIty, as shown III figure 2. The data were calc,:!l ated fro~ the straight lines through points on the
pIt depth-tIme curves using rectangular coordinates.
There now appears to b e a definite trend to lesser
pitting rates as resistivity increases. The effect of
pH of. the soil on maximum penetration rates is
ShOWl~ III fig;ure 3. Note that ~he majority of very
COITOSlve SOlIs 500 n-cm) are III the alkaline ran cre
(> pH7).
0
The effect on maximum penetration of increasin o'
~he area exposed from tha.t of a specimen (about
ft2) to a l arger area (eqUIvalent to t hat of a 20 ft
length of 8 in. diam pipe) exposed to identical soil
conditions, by cfl1culations previously described is
~ hown in figure 4. D ata pertaining to t he effect of
Illcreased area are tabulated in table 1.
. The estimated eJ!ect of increa~in g exposure time
from 5 to 30 years IS also shown 111 table 1 as based
on linear extrapolation of data plotted ' on both
rectangular and logarithmic coordinates. As would
b~ expected, t~e pit depths extrapolated from log
pIt depth-log tune curves are for t he majority of
soils somewhat smaller than the values from rectangular coordinates: Previous investigators, [7]
who se work resulted 11l eq (3), favored t he loo'arithmic relationship and their conclusions seem 0 to be
reasonable.

0.4

V>

E 1000

.:!2100 0

vi

a::
>-

..

It)

f-

<t

100

f<l.

I ena::

........

... ...,."i.. .- . : ,......,..

. .. .
.... .

f<J1
W
<l.
W
W
0

'E

f-

<t

10

10 '

10'

f<J1
W
<l.
W
0

10 '

100

f<l.

10

-... ,.

to

....

>-

..-.-,.... ,.-, ...... .....,.

- C.

10
10

10'

10'

10'

105

10 '

RESISTIVITY, .a-em

RE SIS T IVI T Y, .a -em

4. !II aximu1n pit depths at 5 yeaTS of unMTgl'ound

ex posure calculated Jor lal'gcr areas (e qu'';valent to 20 f t of 8 in.
diwn bare pi pe) /1"O'In data on .small pip e specimen s (fig . 1) .

FIG U RE

FIGUR E 1.
Avera(le of max'';mum pit depths on fTom 6 to 14
f errous specimens bltried in each of 86 underground soi l sites.

Pit de ptlls are take n from the pit depth versus time curves at 5 years.

10'
C!!

10 0

......

<J)

a::
>-

10

E
<t

- .,-'

.. - -
-. ,...

.-.
- .-.. - ..

It)

w 10'
>-

r-

a::
w

lL

<t

f-

r-

<t

a::

i.

<l.

X-

<t
:::2'

.1
10

a::

<J)

a.

w
w

10'

10'

100

vi
a::

o
10

>-

'b

co

.,0 ""

. .:,....

fu.

<t
w

'"a::

f-

<!

a::
:i.

SOIL RESISTIVIT Y
500 .Il.-em

.<
o

500 3000

3000- 60 0 00

a.

X
<t

::E
.1

10

12

pH
FIG U R E

JI ax imwn rates of pen etration (fig. 2) on specim ens

as related to pH for 3 ranges of soiil'esistivity,

3.

7 53 - J S O~65 --~G

5. S am e as fig u re 4, except that the exposure time is

D ata adjusted for the larger ar ea shown in t he

last column table 1, are plotted against soil resistivity
fig ure 5. Perforation of a pipe wall, r anging in
t hickness from 0.172 in. to 0.322 in. for 8 in. diam
steel pipe, is predicted in almost all soils h aving
resisti vities less than 1000 D-cm. In some of t hese
soils, many perforations would probably occ ur. For
soils with resistivit ies over 2000 rl-cm , a nd assuming
t he absence of stray currents or co ntact with more
noble metals, the extrapolated data indicate t hat
perforation of 8 in. diam (0. 322 in. wall-t hickness)
steel pipe in 30 years is r ather unlikely. However ,
the data definitely show t he need for protective
meaSID'es, such as coatin gs, cathodic protection or
both, on wrou ght materials exposed to soils with
resistivities less t han 2000 rl-cm and also even in
some soils of higher resistivity, as indicated by
predicted pit dept hs in a few of the soils shown
abo ve 25,000 rl-cm, all depending on t h e hazard
involved should a perforation occm.
R ecen t work by Scott [3, 9] has a direct bearing
on the relation between soil resistivity and maximum
penetration. Based on random measmements of
soil resistivity in the field h e found that a plot on
log-probability coordinates of soil resistivity measurements for a given area versus estimated cumulative
probability assumes a linear relationship in many

2. Rates of penetration oJ the specimen s (fig. 1), based

on the maximum p'';t depth versus time curves, f rom 5 yew 's to
fTom 12 to 17 yew 's (Jor the majority of the soils),

'f'

10'

10 '

FIGURE

'>-

10'

extm polated to 30 years .

10 '

RE SIS T IVI T Y, .a - em

"-

1
10 '

R ESISTI VIT Y, D.-e m

F I G UHE

10'

f-

10
10

<t

lO' -

. ...

ff-

.... .-........ -.

.- . -" . ,e. .. _...

..... ..::,. ,. -....
... -:..':.

".,

~ e

f-

..

vi
a::

.:!2
E

73

TABLE

Maximum pit-depth '

1.

Time- 5 yrs
Soil site

No.

p'

pH

Interna l
drainage d

fl-cm

1
2
3
4
5
6
7
8
9
10
11
12
13
14

15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30

:ll

32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
51
53
55

56
58
59
60
61
62
63
64
65

66

1215
684
30000
6670
1345
45100
2120
350
2820
7460
11000
3190
290
3520
489
8290
5980
1410
) 970
2870
2370
5) 50
278
11400
1780
2980
570
408
1270
1300
20500
5700
800
4900
2060
11200
11200
38600
7440
970
1320
13700
60
1000
2(i3
1500
1770
190
17790
5210
406
712
1660
218
943
6920
84
62
148
232

7.0
7.3
5.2
5.6
7.0
5.9
4.4
7.6
6.8
6.6
5.3
7.1
9.5
6.2
7.5
4.4
4. 5
7. 3
4.6
7.5
6.2
4.9
9.4
4.5
7.2
7.3
6.6
6. 8
4.2
7.0
4. 7
7.3
6.8
6.7
7. 3
4.5
3.8
4.5
5.6
6.0
5.5
4.7
3.1
5.8
7.4
7. 0
7.6
6.2
4.8
5.8
7. 1
4.8
5. 6
2.6
6.8
4. 5
6.9
7.5
8.0
8.0

P
P

P
P

G
G
F
F

P
F
F

G
G

P
F

G
F

G
J<'
G

VP
VP
VP
P

G
G

VP
F

G
G

G
J<'

F
l'

VP

G
P

VP

G
G

VP
VP
VP
P
P
J<'

VP

J<'
G
l'

A ssumed

Mils

Mils

48
44
68
34
40
20
27
54
35
41
64
50
65
80
52
56
31
43
48
26
56
56
105
20
42
60
33
83
73
27
33
37
43
33
22
42
53
26
34
58
50
79
80
55
40
75
17
131
58
62
66
57
28
40
44
60
40
134
82
H2

96
82
134
72
79
39
47
100
79
98
124
97
161
169
83
98
56
72
77
63
103
83
202
45
63
114
74
137
169
48
75
81

234
8.8
VP
110
886
7. 0
P
26
Ie
261
7.3
104
]e
103
7.3
88
}'
81
7.3
97
8500
4.6
G
84
28000
4.8
G
51
25000
4. 8
G
63
54000
4.8
G
74
44400
4.8
G
68
497
8. 4
P
83
531
10.2
95
P
51
7. 3
P
102
149
7.4
P
185
Ie
102
7.4
177
320
7.1
VP
89
3450
6.9
G
70
320
9. 2
F
107
106
8.5
P
123
273
7. 3
F
160
10800
4.7
G
70
16400
4. 8
G
22
Ie
122
552
7.4
52
) 23
6840
4.1
P
45
124
1160
4.4
P
62
125
5770
3.9
P
52
Average of the deepest PIts on 6 to 14 speCImens (a rea, a pproxImately 0.4 fF)
for each soil.
b Romanoff [1] , table 6.
52
54
101
102
103
104
105
106
107
108
109
110
III
112
113
114
lL\
116
117
118
119
]21

e
d

Soil resistivity, saturated soil at 60 F.

Also aeration; 0, goo d; F, fai r; P, poor; VP , very poor.

Romanoff [1], fro m pi t depthtime c ur ves u sing data in tables 13, 14. 15. There
were from 6 to 14 specimens involved in each removal from a given soil and from
3 to 6 removals, the 3 r emovals being for t he 100 series of soils.
o

'rime- 3~

yrs

Specimen
Specimen
area Al e

areaA 2

9.)

69
99
67
131
54
69
115
78
125
201
HI
78
181
52
266
118
126
134
H6
57
81
89
122
81
272
166
228

pitt ing rate

a fter 5 yr s g

S pecimen

rect.angular h
co-ord

Mils /vr
6.0
1.6
1.3
6.7
1.4
0.28
2.2
4.4
2. 1
1.0
1. 5
1.3
6. 0
3.2
0.84
2.3
1.3
1.1
2.7
4.4
4. 4
1.6
8.3
0.75
1.0
0.61
3.0
17.0
8. 1
3.2
1.3
2.7
8. 4
4.1
0.60
1.0
3.2
0.7
5.4
8.1
4. 7
4.0
12.2
2.5
7.3
4.2
2. 0
13.6
1.4
1.3
15.0
12.6
6.0
5.0
4.3
3.0
6.7
5.7
6. 9
6.0

l\lils
198
84
101
202
75
27
82
164
88
66
102
83
215
160
73
114
64
71

116
136
110
96
312
39
67
77
108
508
276
107
66
105
253
136
37
67
133
44
169
261
168
179
285
118
223
180
67
471
93
95
441
372
178
165
152
135
208
2i7
255
262

Specim en

log.co-o rd

lV il.
190
72
97
165
86
26
76
130
82
70
90
75
125
165
84
100
57
76
92
115
120
85
310
32
65
67
100
420
235
95
69
81
270
92

43
58
103
48
140
185
)30
135
270
95
240
140
60
550
87
90
390
305
260
145
160
HO
155
240
195
200

Assum ed
areaA2 j

Mils
378
134
192
351
170
52
131
244
185
167
174
]4.\

311
348
134
175
103
128
148
278
220
125
597
72

98
118
223
693
545
170
156
177

598
J93
J93
92
254
99
284
365
203
214
678
191
469
337
183
1117
177
183
792
619
528
294
325
223
314
487
396
406

873
223
18. 0
560
430
203
154
100
53
5.1
274
2H
3.0
179
135
376
185
4.
4
198
179
420
853
197
17.9
644
244
120
171
2.8
154
117
0.9
74
58
104
199
98
123
128
2. 4
294
182
150
4.3
145
203
2.2
123
100
138
355
218
175
169
5.4
609
300
193
H.7
388
650
11. 8
397
320
207
2530
1250
35.0
1060
375
2030
35.0
1052
1000
360
426
210
5.1
217
181
244
142
1.8
115
120
670
330
217
13.7
450
437
215
7.6
313
250
1240
795
610
25. 4
325
189
93
142
1.7
113
28
.\7
0.4
32
45
77
156
2. 4
112
106
78
158
1. 6
85
91
233
152
115
3.6
126
223
3. 1
110
106
130
r A 2= 45.16 ft', eq m va lent to a 20 ft length of 8 ill. pIpe . bOe eq (2) . Fo r soils
!through 47, u sed ex ponents a, from Logan. Ewing, and Denison [7]. For a!lother
soils, u sed exponent a=0. 15, from Schwerdtfeger (8).
g Based on best straig h t line after 5 years on pit depth-time curves on linear
coordinates.
h Extrapolated pi t depth-time curves.
i Extrapolated pit dept h-tim e curves.
j From log-coordina tes (i) and u se of eq (2).
~

74

I
1

are,Ls a nd is therefore indicative of a uniform environmen t. LaLer, Scott offered a theoret ical basis for
his empirical relationship [10]. Soil resistivity
m etts urements made by him along pipeline rig hts-ofway together with leak r ecords on th e pipelines show
that leaks occur in the places of low resis tivity and
alm ost invariably at resisti lTit ies below the value
shown by 50 percent probability on t he log r esistivity-prob ability curve.

TABLE

2.

Corrosion currents associated with weight losses o '

specimens a
Corrosion rate

Soil
site
N o. b

p '

-----0- 2 y r

After 5
yr e

-------- ---

0- 2 y r d
N O, b
After 5
yr e
- - - - - - - - - --

mA /ft '
2.0
1.1
0.41
1. 3
1.1

51
53
55
56
58

190
17i90
5210
406
712

mA/fl'

mA/fl '

45100
2120
:350
2820
74(;()

0.28
1.1
1.8
23
2.9

0. 26
0. 59
1. 8
1. 0
1.1

59

(i3

J660
218
9] 3
6920
84

2.5
8. 4
3. 4
5.3
4. 7

1.6
5.6
1.6
1.1
2.5

12
13
14
15

11000
3 190
290
3520
489

0.90
l. 3
5. ()
l. 7
3. G

0.34
0.8 1
:l.6
l.1
1. 8

64
65
6G
52
54

62
148
232
Zl4
886

20.3
9.4
12.5
(i. 2
1. 5

12.7
1.4
2.8
6.2
1.1

16
L7
J8
I!!
20

8290
5890
141 0
1970
2870

3 G
:l. 1
1.2
2.0
2. 9

1. 7
1.1
0.56
0.5:1
l. 5

101
102
103
I(}I
105

261
103
81
8500
28000

7.0
7.2
6.9
4.6
.5.3

f) . Ii

21
22
2:3
24
25

2370
5 L50
278
11400
li80

4. I
2. 9
J I. 2
062
1. :3

1. 5
0.72
4.7
0. 22
0. 65

106
107
108
J09
11 0

25000
54000
44400
497
53 t

:1. 9
:3. G
5.2
8.7
6. 9

26
27
28
2n
:l0

2980
408
1270
L300

2. 2
2.0
5.9
5.9
2.8

O. (lS
l. (i
5.5
4.2
1. 2

III
LI 2
11 3
J 14
115

51
149
102
:320
:1450

9.2
11. 9
J3.5
4.8
2.8

5. !)
7.5
9. !I
:1.8
0.411

31
:12
33
34
35

20500
5700
SOO
4900
2OGO

1. 6
1. :3
:1. 4
2.9
1. :3

0.50
0.84
3. 4
1.1
0.1 2

11 (;
LL 7
LL S
119
J2 L

320
106
273
10SOO
1&100

10. I
J2.8
8.7
2. S
1. 4

4. (l

:16
:17

1. 7
3.7
O. G2
2.7
3. 4

0. 40
1. 9
0.53
1.2
I.G

122
IZ1
124
125

552
6840
II (;()
5770

3. I
4.8
9.0
5. ()

1.4
2. 0
I. -I
1. :1

40

11 200
11 200
:)S(iOO
7410
970

41
42
43
44
45

1:320
13700
GO
1000
263

2.0
4.5
4.8
I. L
3. I

0. 91
3.0
:1.9
1. 0
2. S

46
47

1500
L?iO

2. :1
1.0

20
o 97

5
6
7

75

p '

mA /fl'

The weight losses on th e speci mens for which

maximum penetration has been g ive n were co n ITer ted
to corrosion curren t densities by applying Faraday'S
law a nd the usual ass ump tions that th e iron was
oxidized to the di valen t state a nd t h e corrosion
effi ciency was 100 percent. Th e data tab ulated in
tlLble 2 are based on weigh t losses measured at 2-year
removal time ftnd on remo vals afte r 5 years. Th e
co rros ion rates, expressed as current densities, plotted
versu s soil resis ti vity ,1nd pH ar e show n in fi gures 6,
7, and 8. The current den sitie.s calculated from t he
weight losses of pecimens based on th e fLrs t rem ovals
(ttpproximately 2 years for m ost specim ens) are
s hown plotted versus soil resistivity in fig ure 6.
Wha t was previou sly said abou t Lhe sig nifica nce 01"
500 Q-cm as a di vidin g lin e I" or difrerences in maximum penetration dmin g e,U"ly ex posure also seems
to apply to the currellt densities associated with
corrosion. The effect of soil resis tivity in substa ntially reducing corrosion curren t den si ties for exposur e
beyond 5 years is illustrated by figu re 7. These
data a re t he r esult of straight lin es on rectang ular
coo rdin ates drawn through the weigh t losses a fter 5
years on the weight loss-time curves (not sh own).
The sa me data as related to soil pH a re shown in
fi g ure 8.
Th ese corrosion currents are primarily of in teres L
ill conn ection wit h current den sities necessar y for
caL hodic protection. Th e relations ilip between "<llu es
01" corrosion current a nd cunent required for cat hocli (;
protection dep ends on the type of control of t he
cOlTosiol1 r eaction. Theoretically , if th e reaction is
under strict cathodic cont rol (no anodic polarization ), t he protective curren t is eqmtl in magnitude to
tile co rrosion current. This ideal situ ation is not
us ually realized with ferrous metals used underground , the protective current requirement bein g
somewhat greater. For fmther information on this
rel atio nship and som e d ata, the reader is r eferred to
p a pers by the author [11 ,12] and other investigators
[13, 14,15]. D ata obtain ed on steel pipe specimens
exposed und erground sh owed that the r atio, I 1,/io, of
protectilTe current, I p , to corrosion current, io, varied
between 1 a nd 2, depending Oil the length of tim e the
specim ens were exposed and on t he chemical and
physical properties of t he soil . For example, in the
case of a specimen b uri ed in a soil of 7500 Q-cm
resistivity, t he ratio W,lS abo ut 1.3 during the first
week of buri:ll a nd aroun d 2.0 at 6 months and
t hereafLer for 16 m ont bs exposure when t he specimen
was Iemoved [16] . For a similar specimen (area,

site

f!-cm
12 15
(lS4
30000
6670
1345

1
2
3

4.2 . Relation Between Weight Losses and Currents

Required for Cathodic Protection

Corrosion rate
Soil

9
10
11

as

:19

570

3. i
3. 1
2.8
3. I
1. 8

fl -cm

(;()

61
62

14.2
5.0
4. 4
6.5
5.2

3.6
0.50
041
6.5
3. 4

2.2
f), li
I. .)

o.:n
1.!)

l. 0
O. .I!!
2. (;
n.2

5. :~
5. :{
1. 8

o. IU

!~ Calculat ed from we i ght lossLimc curves on data b y ilom an ofr II I t1 s ill ~ I i'ara~
d a.y's law, assumin g 100 pcrcent co rro sion eni ciency a nd Lha. t th e iron is cti va lent.
For weight losses, sec Roma ll oO' 11] , t a.b les J3, 14, Hnd 15.
b Rom an off [11 . t able 6.
e Soil reSistivit y, sa turated soil at. GO OF .
d Based on first relll ova l from soil , u su a ll y 1 to 2 years.
e Ba sed on the best stra ight line on the wrig ht loss-tim e cur ves afiel' 5 years.

0.4 f t2) buried in a. tidal il1,ush (3 00 Q-cm, r esist.ivity)

for a period of 32 mont hs, t he corrosion reaction WItS
under co mplete eat hodic co ntr ol durin g the fLl"st week
of exposure; the I p/io ratio W,lS 1.3 at 8 months, a nd
abo ut 1.6 at 32 mont hs when t he specimen was
removed (fig. 9) . The r aLios are believed to be
a pproximately correct becttllse a fter the specimens
were r emoved (rom t he trenches they were cleaned,
weighed and the ,tctual weigh t losses were in reasonably good agreeme nt with th e weight losses calculated
from t he polarization data. Thus, after 32 months
(fLg. 9) t he curr ent density require~ for cathodic
protection (based on I p , area, 0.4 ft2) was about
8.2 mA/ft2. Similar polarization d ata on steel
specimens of the same size were obtained by Ardahl

100

......

N_

::::<l

.
.. ..........
"I.... .
.
..,'
.::..... . ...

r-:

10

3 rd

a::
a::

.......

::>
0

Z
0

e...... :

a::
a::

, I

10 '

, I

la'

10 '

10 '

6. Corrosion CU1Tent den sities calcu lated jrom the

weight losses oj the s pecimens (fig. 1) ajter removal jrom the
underground sites u pon 2 years oj exposw e .

FIG U RE

100
N

"<l

10

w
a::
a::

::>
0

.....-._:
I .. :

e. .. .,_

., .. . .- ...
....

. e.!.

ifi
0

a::

a::
0
0

_:.~

.1
10'

10

10'

10 '

10'

RESISTIVITY. D-e m

7. Oalculated corrosion CWTent den sities based on the

rates of weight loss between 5 and 12, and between 5 and 17
years of eXpOSlt1'e f0 1' the maj ority of soi ls.

FIGU R E

SOIL RESISTIVIT Y

:: 12

' <500 n-em

<l

"-

E 10 ~:

500 3000
3000- 6 0 0 00

w 8
a:
a:
(3 6

.'t

z
Q 4
en
o
a:
a: 2
o

"

10

10

12

14

16
~

FIG U R E

APP LIE D CURRENT, mA

RE SIST IVI T Y, D -em

f--'

:r--o--

MONTH

3 2 nd MONTH
C.R.'S.l mA / ft 2

IP

10

C.R.'6.S m A / ft 2

.1

8t h

..

I -

ifi

DA Y

G. R., 2.5 m A /f t 2

II

12

9. Polal'i zati on curves on a plain steel specimen bu ried

1m derground jor 32 months in a tidal marsh.

o anodic

. cathodi c

on soil properties and on duration of exposure. In

a laboratory study, [18] a steel tube (area, 24 in.2) was
exposed to a 20,000 !I-cm sandy loam soil for 60
days. The current density necessary for cathodic
protection was found to be about 3 times the value
of current density associated with the weigh t losses
of similar tubes which were permitted to corrode
freely, that is, the 1 p /iD ratio was about 3.
On the basis of the foregoing 1p/io ratios , the corrosion current densities given in table 2 and in figures
6 and 7 might be increased accordingly and thus
converted to values necessary for cathodic protection .
This does not mean that current densities are substituted for polarization requirements as criteria [11]
for cathodic protection. The protective current
densities are presented to aid in estimating total
current requirements in the design of underground
cathodic protection systems. It would seem reasonable to multiply the corrosion current densities by
1.5 for soils having resistivities between 0 and 5000
!I-cm . A factor of 2.0 might be used for soils with
resistivities between 5000 and 10,000 !I-cm. Above
10,000 !I-cm, the 1p/io ratio could be taken as being
3.0 and still result in relatively small protective
current densities for most high resistivity soils.
A comparison of the corrosion CUlTent densities in
the two columns (table 2) and of Figures 6 and 7
shows that it would be economically advisable, at
least for uncoated underground structures, to allow
about 2 years before designing the electrical requirements for cathodic protection. The current densities
necessary for cathodic protection are considerably
reduced after a few years of exposure, especially in
the soils of high resistivity.

pH

FI GURE

8.

5. Conclusions

Same as fi gu re 7, exce pt as re lated to pH for 3 ranges

of soil resistivit y.

The maximum pit depths on commonly used

wrought ferrous materials bmied for periods up to
5 years in back-filled trenches are on the average
deeper in soils with resistivities below 500 !I-cm than
in soils having higher resistivities while in the soils
above 500 ohm-cm, there appears to be no regular
pattern between maximum pit depth and soil resistivity. However, for periods of expos me in excess

[17]. The writer analyzed these data and observed

that after 2 years of underground exposure, in 2 soils
with resistivities between 550 and 1000 !I-cm, the
1pjio ratio was between 1.1 and 1.8.
In higher resistivity soils (great.er than 10,000 !I-cm) ,
the 1p/io ratio is probably about 2 or more, depending

76

3. Average value8 of maximum pen etration rate and

con'osion rate (current) on specim en8 jar 3 mnges oj soil
resistivity after 5 years of exposure

TABLE

Maximum penetration
Soil resistivity

n-cm
50 to 500 .........
500 to 3000 .......
3000 to 54,000.....

N umber
of soil
sites

ra.te

Min

Corrosion
current b

Avg

Max

Min

Avg

,"lax

mils/yr
26
28

32

60
0.. 84
.28

4. 2 1
1 11.4
2.2

35.0
12. G
G.7

1.. 12
4
. 19

1m1./t
I. 6 1

12.7
6.2

0.99

3. 0

Calcu lated from t he specimen pittiug rates after 5 years (table 1).
b Calculated from t he corrosion rates (currents) after 5 yea rs (table 2).

of 5 years, the rate of maximum penetration lesse ns

as t.he soil resistivity in creases heyond 500 Q-cm.
This is shown in table 3 wh ere data from 86 soil sites
are averaged for 3 ranges of soil resistivity. Based
on an empirical equa tion , increasing the area exposed to that equival ent to a 20 ft length of 8 in.
dia m pipe (45. 16 ft 2) appro)"'i mately doubles the maximum penetr ation rates shown in th e table which
apply to the actual specimens (0.4 ft2) .
In soils with r esistivities up to 500 Q-cm, perforation of thick-wall (0.322 in. ) bare wrought ferrous
pipe is predicted in tbemajority of soils a fter 15
year s of exposure. After 30 years of exposure, the
same would be true probably in the majority of
soils having resistiviti es to 1000 Q-cm; in some of
t hese soils many perforations would be likely Lo
occur.
After the corrosion rates of the ferrous specimens
became fairly well s tabilized as shown by weight
loss-tim e curves, the weight losses were co nverted
to current densities. Corrosion current den si ties,
a\Ter aged for t hree ran ges of soil resistivity, are given
in table 3. The current densities r eq uired for cathodic pro tection are greater th an the curren t densities
associated with the corrosio n. By anticipating the
type of corrosion con trol p eculi ar to the cOlTosi ve
en vironment, corrosio n curren t densities can be adj usted to curren t densities necessary for cathodic
protection. A factor of 1.5 is suggested for soils wit h
resisti vities up to 5000 Q-cm, a factor of 2.0 for soils
from 5000 to 10,000 D-cm, and a factor of 3.0 for
soils with resistivities above 10,000 Q-cm. It is
considered to be economically advisable to wait for
abou t 2 years of exposure before measuring the curren t densities required [or the cathodic protection of
bare II ndergrollnd structures.

6 . References
[I] Melvin Romanoff, Underground corrosion, N BS circ.
579, U.S. Government Printing Office, Washington ,
D. C. (1957).
[2] Kirk H. Loga n, E ngin ee ring significance of National
Bureau of Standards soil-corrosion data , J . R es.
NBS 22, 109 (1939) RP1l71.
[3] Go rdon N. Scott, The di stribu tion of soil conductivities
and so me consequences, Corrosion 140, 396t (August
1955) .
[4] Edwa rd Schaschl an d Glenn A. Marsh, Some New
Views on So il Corrosiotl, :Materials Protection 2,
8 (Nove mber 1963) .
[5] Mcl vin Romanoff, Co rrosio n of steel pilings in soils, J.
Res. N BS 66C (Eng. and In str. ) No. 3,223 (July-Sep t .
1962).
[6] G. N . Scott, Adjustment of so il-co rrosion pit d ep th
meas ure ments for s ize of sa mpl e, Proc. American
Petroleum Instit ute 140, No.4, 204 (1933).
[7] Ie H . Loga n, S. P . Ewing, and 1. A. D enison, Soil
co rros ion test in g, Sy mp osium on Corrosion Testin g
Procedures, AST1VI (J 937).
[S] ' V. J. Sc hwerd tfeger, Laboratory meas ure men t of the
co rros ion of fel'l'ous metals in SOils, J. Res. N BS 50,
329 (June 1953) RP2422.
[9] Gordon N. Scott, Dist ribu t ion of soil conductivity an d
its relation to underg round corros ion , J ournal A W'V A
52, 37S (Ma rch 1960) .
[10] Gordon N . Scott, A t heory for t he probability distribution of soil cond uctivi ty, COl'l'osion 18, 25lt (Jul y
1962).
[11] W. J . Schwerd tfeger and O. N. McDor rnan, Po tential
and current req uirements for t he cathodic protection
of steel in so ils, J . R es. NBS 407, 104 (Au gust 195 1)
RP2233 ; Co rrosion 8, 391 (1952).
[12] W . J . Sc hwe rd tfeger and O. N. Mc Do rman, :Meas urement of t he cor rosion ra te of a metal from its polarizin g characteristics, J . Electro che m. Soc. 99, 407
(1952) .
[l 3] 1. A. D enison, Electrolytic bc' hav ior of ferrous and
non ferrous metals in so il-co rros ion circui ts, Trans.
Electrochem. Soc. 81,4:35 (1942).
[14] Robe rt Pope, Cell curren ts and poten tials, Cor rosion 11,
IS9t (April 1955).
[15] E. Schasch l a nd G. A. Ma rs h, The effect of dissol ved
oxygen on co rrosion of steel and on c urren t req uired
for catho d ic protection, Co rrosion 13, 243t (Apl'iI1957).
[16] 'vV. J . Schwerdtfeger, A Study by polarization techn iques
of the co rrosion rates of alumin um and steel und ergro und for sixteen mont hs, J . R es . NBS 65C (Eng.
and In str.) No.4, 271 (Oct.- D ec. 196 1).
[17] E. II . Ardahl, Corps of Enginee rs, U.S. Army, Lower
Mississ ippi Vall ey Di vis ion , Vicksb urg, M iss. (private
co rrespondence) .
[IS] 'V . J . Schwerd tfege r, Current and potential relations for
t he cathodic protection of steel in a high resistivity
environment, J. R es. NBS 63C, 37 (Jul y- Sept. 1959);
Corrosion 16, 209t (May 1960).

(P aper 69Cl- 188)

77