Old Bridges Give Clues to Steel Deck Performance

ROMAN WOLCHUK

This paper was presented at the AISC National Engineering Conference, Omaha, Nebr. in May 7964.

W H E N E V E R STEEL DECK bridges are discussed, the ques- Figure 1 shows the cross section of the 6-lane Harlem
tion invariably arises: How well do the steel decks per- River Bridge, a 310-ft lift span.
form over the years? More specifically: H o w safe is a T h e cross section of the Bronx-Kill Bridge, a 350-ft
steel deck against corrosion? H o w good must the wearing fixed span structure, is very similar, except that the
surface be to protect it adequately? Is there a danger that bridge is wider and has eight lanes.
the steel deck, with the top surface normally not ac-
cessible, will deteriorate under the surfacing? ORIGINAL SURFACING
T h e answers to these questions are provided by Both bridges were originally paved with 1-in. thick
actual experiences with steel deck bridges. mineral-surfaced asphalt planks, 24 in. long and 12 in.
T h e oldest of the new ("orthotropic p l a t e " ) type steel wide. T h e steel deck was painted with red lead and the
bridge decks in Europe are, by now, 15 years old. Some planks were bonded to the deck with asphalt cement.
steel deck structures of older types are more than twice More complete details of the original plank surfacing
that age, and considerable experience with them has are given in Section 8.3 of the A I S C manual for steel
been accumulated. plate deck bridges. 2
In this country, steel deck bridges of the new kind Performance of the planking on the H a r l e m River
are still in the design or pre-construction stage. However, Bridge was quite good, and at the time of repaving
there are many "battledeck floor" type bridge structures about 70 percent of the deck was still covered by the
in service, built in the 1930's. original plank. However, performance of the Bronx span
T w o notable examples are the Harlem River Bridge planking was definitely less adequate, and almost all of
and the Bronx-Kill Bridge in New York City built in the planking had been replaced during the 27 years.
1936 and recently resurfaced. T h e writer had an op-
portunity, on behalf of American Iron and Steel In-
stitute, to investigate the conditions of these steel decks
after 27 years of service.
T h e observations made 1 offer some clues to the ques-
tions raised.
Both structures are located on the approaches to the 3'-4"
New York City Triborough Bridge; one of them (the
SO'-e*' -4
1" surfacing
H a r l e m River Bridge) provides connection to the
'R(s)
Franklin D. Roosevelt Drive and the local streets in M a n -
X-7I20@1'-2"
h a t t a n , the other connects with the Major Deegan a n d
Bruckner expressways. Traffic on both bridges is ex-
tremely heavy with the difference that the Harlem struc-
ture carries few trucks, while the Bronx span has a very
27CB98 ' A2I35 @T-0"
large a m o u n t of truck traffic.
Fl. beams @ 2 5 - 1 0 "

-U-8'-7'V 37'-9"
Roman Wolchuk is a Consulting Engineer, New York, N: Y.

1. Report on Repaying of the Harlem River Bridge and the Fig. 7. Cross section of Harlem River Bridge
Bronx-Kill Bridge {unpublished), submitted to American Iron
and Steel Institute by Roman Wolchuk, Consulting Engineer, New 2. Design Manual for Orthotropic Steel Plate Deck Bridges,
York, September, 7963. American Institute of Steel Construction, New York, N. Y., 7963:

137
OCTOBER/ 1964
T h e r e are indications that the difference in performance rosion effects were found, were confined to these narrow
of the planking on the two bridges was due, primarily, strips along the edges of the plank.
to the difference in the quality of the original plank
and the difference in method of placement of the planks CONDITION OF THE DECKS
on the deck. After removal of the planks, lumps of asphalt sticking to
T h e manner of placement of the plank is responsible the deck were scraped, and rust scale, generally present
for its bond to the deck and is very important because under the plank joints, was removed with air operated
deterioration starts as the bond along the plank periphery wire brushes. Most of the deck area, after cleaning, was
is broken, and the plank begins to crumble at the edges, still covered by the original red lead paint, occasionally
progressively exposing the deck (Fig. 2). overlaid with a coating of asphalt cement.
Regarding deck protection provided, it should be Typical conditions of the deck are shown in Figs.
noted that asphalt planking even in its original condi- 4 and 5.
tion cannot prevent access of moisture to the deck at T h e effects of corrosion consisted of streaks of shal-
the joints between the plank. Although the Triborough low pock marks, pits and scars following the outlines
Bridge and Tunnel Authority, owner of the bridges, of the planks where the bond between the deck and the
does not use snow melting salts on its structures, it can plank had failed. Outside of these strips, generally from
be assumed that the decks are not free from contamina- 2 to 4 in. wide, the deck was not affected by corrosion.
tion by salts in wintertime, since salts are used freely by Depths of corrosive pits and scars in the deck surface
the city of New York on the approach expressways. ranged in typical cases from 4 to 40 mils. Depths were
T h e decision by the Triborough Bridge and Tunnel measured directly by a dial indicator which could be
Authority to repave both bridges with a 1-in. thick moved horizontally along a steel base. Fig. 7a shows a
course of asphalt concrete rather than with asphalt characteristic profile across the affected strip corre-
plank was motivated primarily by the high cost of plank- sponding to the conditions shown in Fig. 4; Fig. 7b
ing (about three times that of the asphalt concrete) and shows a similar profile corresponding to the somewhat
the fact that high-quality paving plank, to match the deeper scars in Fig. 5.
quality of the original paving, could not be supplied by Deeper and more extensive pits were observed only
the manufacturers. infrequently. Figure 6 shows a cluster of pits and scars,
surrounded by an area entirely unaffected by corrosion;
PLANK REMOVAL AND DECK CLEANING the corresponding cross section is given in Fig. 7c. It
Repaving operations started with removal of the plank- shows that the depth of pit measured from the original
ing, using a 7 ton front end loader and manual labor. steel deck surface was about 75 mils, or somewhat less
Adhesion of the plank to the deck was very strong, than 3^6 m - T h e deepest pit recorded anywhere on the
and the bond was unbroken under the middle portions decks was 115 mils. In general, however, conditions shown
of the planks (Fig. 3), where the original asphalt cement in Figs. 8a and 8b were typical.
was found to be still tacky. Around the periphery of the In addition to dial indicator measurements of the top
planks the bond was broken over a width of 2 to 4 in., surface conditions, total thickness measurements of the
and dust from the pulverized bituminous coating and the deck plate were made in several locations with an Audi-
plank material had accumulated there. Whatever cor- gage ultrasonic instrument.

Fig. 2. Deteriorated original planking, Bronx-Kill Bridge Fig. 3. Removal of planks by pick. Shiny surfaces are areas where
the bond between the plank and the deck was intact

138

A l SC E N G I N E E R I NG JOURNAL
 T h e readings consistently indicated a plate thickness small quantity did not form a continuous membrane, it
in excess of the original nominal thickness of % in. or was believed that a larger amount of bitumen may cause
625 mils and generally ranged from about 635 to 645 instability of the overlaying asphalt concrete.
mils. This excess thickness indicates the usual mill T h e composition of the asphalt concrete mix was as
overrun of a nominal % in. plate. T h e Audigage meas- follows:
urements confirmed that there was no overall loss of
Weight
the specified plate thickness, and that no appreciable
corrosion loss had occurred on the under side of the (%)
1. Stone (crushed trap rock, }/ in. max.
deck, which was not accessible for inspection.
size) 24.3
Thus, the steel decks have been found to be in good
2. Sand (coarse, natural) 58.0
structural condition after 27 years of service.
3. Limestone dust 5.0
THE NEW SURFACING 4. Asbestos fibers 7 M 0 6 (7M Grade by
Quebec Standard screen test, min 6.3
T h e broom-clean deck received a hot asphalt tack coat
percent retained on 10 in. mesh, max.,
(100 percent bitumen 60-70 penetration) of 0.05 to
25 percent passing 65 m,esh) 3.1
0.10 gallons per square yard. Though this relatively
5. Asphalt, 60-70 penetration (Specifica-
tion M 5 of N. Y. State Department of
Public Works) 9.6
100.0%
After an initial adjustment, the approximate ag-
gregate gradation, including the limestone dust and the
asbestos was as follows:

Sieve Size % Passing

H in. 100
#4 90
#10 68
#40 36
#80 14
#200 10

T h e characteristic feature of this mix is the relatively

high bitumen content of 9.6 percent, which tends to
improve the density and impermeability of the pave-
ment. A recent report indicates that such an increase of
Fig. 4. Dial indicator measurements of typical deck surface condition.
Corresponding profile is shown in Fig. 7a

Fig. 5. View of typical deck surface condition, revealing only shallow Fig. 6. Deeper corrosion pits, observed infrequently. Surrounding
corrosion marks and pits. Corresponding profile is shown in Fig. 7b area unaffected by corrosion. Typical profile in Fig. 7c

139
OCTOBER / 1964
 - Original surface of steel plate
no corrosion 0.075 in ^0.055 in
/r . /|Corr. pock marks

-&
0.004 in J
f?

.E o Fig. 7. Typical deck surface profiles

in JJ
CM 0)
(a) Profile corresponding to Fig. 4

(b) Profile corresponding to Fig. 5

(c) Profile corresponding to Fig. 6
SCALES

(a) (b) (c)

asphalt content without sacrifice of stability is possible European steel deck bridges, built both before and after
if asbestos fibers are added to the mix. 3 World W a r I I .
T h e new pavement has a smooth sandpaper-like Such findings may also be of importance in estab-
surface texture. T h e friction value has been found satis- lishing the criteria for the design of wearing surfaces on
factory. T h e pavement has successfully withstood its steel decks.
first winter under traffic. Basically, a wearing surface has to satisfy two re-
quirements: (1) provide a durable, stable and skid-re-
sistant riding surface, and (2) protect the deck against
DISCUSSION corrosion. In the light of these observations, it seems
From the measurements and observations of the condi- that primary attention ought to be directed to the first
tions of the steel decks discussed above it is seen that deck problem, while the second one appears to be much less
corrosion during 27 years of service was minor, and that critical.
the structural strength of the decks was not impaired. Regarding the first objective, a good and lasting bond
It is well to note that the decks were covered with a between the surfacing and the deck is one of the main
plank surfacing that could not and did not prevent the factors determining the durability of the surfacing, as is
penetration of water to the deck at the joints between the evidenced by European experiences.
planks, and that the water may have been contaminated With regard to the second objective, bond is also a
by salts during wintertime. decisive factor. The observations on the Harlem River
O n e of the reasons why little damage to the decks and the Bronx-Kill bridge decks clearly show that
occurred under the surfacing may be that even though where the bond was intact, there was no corrosion. O n
moisture could be absorbed and held by the dust in the the other hand, the fact that even under faulty surfacing
joints, there was not enough fresh atmospheric oxygen corrosion damage was not significant suggests that costly
necessary for corrosion. multilayered wearing surface systems expressly devised
Thus, if the experience with the Harlem River and for protecting decks against corrosion may be super-
Bronx-Kill bridges is any indication, it appears that the fluous, and that simple and less expensive surfacings,
danger of deterioration of steel bridge decks by corrosion well bonded and otherwise equal to the traffic demands,
is not very acute, even if the wearing surface is not per- should suffice.
fect. This view is supported by the experiences with In conclusion, it may be said that the observations
made on the two steel deck bridges offer a reassuring
answer to the question of steel deck performance. This
3. The Performance of Asbestos-Asphalt Pavement Surface
Course with High Asphalt Content J. H. Kietzman, M. W. should give encouragement to the bridge engineers
Blackhurst and J. A. Foxwell, presented at the Meeting of the contemplating the new steel deck system for their struc-
Highway Research Board, January, 1963. tures.

140
AISC ENGINEERING JOURNAL