Sanders Peyton Hale TRB 2006 August 5 2006 revised

Micah Hale, Assistant Professor Department of Civil Engineering University of Arkansas 4190 Bell Engineering Center Fayetteville, AR 72701 Tel: 479 575-2970 Fax: 479 575-7168 micah@uark.

Investigating the Causes of Bridge Deck Cracking

August 1, 2006

Word Count: 7340

Steve W Peyton, Bridge Engineer

Arkansas State Highway and Transportation Department

P.O Box 2261

Little Rock, AR 72203

Tel: (501) 569-2000

Fax: (501)-569-2400

steven.peyton@arkansashighways.com

Chris L Sanders, Graduate Research Assistant

Department of Civil Engineering

University of Arkansas

4190 Bell Engineering Center

Fayetteville, AR 72701

Tel: (479) 575-2970

Fax: (479) 575-7168

cls20@uark.edu

W Micah Hale, Assistant Professor

Department of Civil Engineering

University of Arkansas

4190 Bell Engineering Center

Fayetteville, AR 72701

Tel: (479) 575-2970

Fax: (479) 575-7168

micah@uark.edu

ABSTRACT

A research program conducted by the University of Arkansas along with the

Arkansas State Highway and Transportation Department investigated the concrete

properties, curing regimens, and cracking density of five bridge decks cast in

Arkansas Concrete was sampled from five bridge decks under construction in

the summer of 2005 The fresh and hardened concrete properties were measured

for all the bridge decks The curing regimens for the bridge decks were also

documented After construction, the research team returned to each bridge to

investigate and measure the amount of cracking The findings from the research

showed that there were many factors that led to bridge deck cracking and there

was difficulty determining if there was a single factor that led to the cracking

Some of those factors included low one day compressive strengths, delayed

curing, and deck movement during construction

Many factors contribute to cracking in concrete bridge decks Some of these factors include structural design, material properties, mixture proportioning, and construction and curing practices During the summer of 2005, five bridge decks were examined to determine which of these factors contributed to bridge deck cracking The research focused on the construction practices, curing regimens, and concrete properties

BACKGROUND

Permeability, durability, and compressive strength are three concrete properties that play

a significant role in the overall bridge deck performance All three of these properties are affected by concrete mixture proportioning and construction practices These concrete properties have been proven as good indicators of bridge deck concrete performance

Permeability is the ability of concrete to resist penetration by water or other

substances Chloride penetration is a major concern in bridge decks (1) Low

permeability concrete can generally be obtained through a low water to cementious material ratio (w/cm) Much research has found that permeability was proportional to w/cm Low permeability concrete provides greater protection against reinforcement

corrosion (2) Decks that had high permeability also experienced severe cracking;

therefore, permeability might be used as an indicator for predicting the cracking potential

of concrete (3) NCHRP Synthesis 333 recommends bridge deck concrete to have

permeability, per AASHTO T 277, in the range of 1,500 and 2,500 coulombs to enhance the performance of bridge deck concrete(1).

Concrete durability is the ability to resist weathering action, chemical attack, abrasion, and any other process of deterioration ACI Committee 201 (Guide to Durable Concrete) recommends that bridge decks exposed to deicing salts have a maximum w/cm

of 0.45 and an average air content of 6% for a nominal maximum size aggregate of one

inch (2) NCHRP Synthesis 333 recommends the use of concrete with w/cm between 0.40 and 0.45 to enhance the bridge deck performance (1).

As compressive strength increases, creep decreases at a higher rate than the rate of increase of tensile strength This is one of the reasons high strength concretes, which

have higher tensile strengths than regular concrete, experience more cracking (3,4).

Cracking tends to increase with compressive strengths, which is most likely related to increases in cement content and paste volume This has been shown to be a direct

relationship in separate comparisons (5) Many agencies have suggested that the trend of

increasing 28-day compressive strengths has led to increased cracking Some specified mixtures can achieve 28-day strengths in three to seven days Compressive strengths for bridge decks should be based on later age compressive strengths, such as 56- or 90-day compressive strength so low heat of hydration cement and supplementary cementing materials can be incorporated into bridge decks without violating strength requirements

(6) Also, early age strengths of concrete should be controlled carefully in order to avoid early deck cracking (7) Finer modern cements typically have one-day compressive

strengths near 45 percent of the 28-day strengths Cement manufactured in the

mid-1940s had one-day compressive strengths of only 11 percent of the 28-day strength (6).

AASHTO and AHTD require a minimum 28-day compressive strength of 4,000 psi for

bridge deck concrete (8,9).

TESTING PROGRAM

The research team sampled concrete from five bridge decks in the state of Arkansas from June 2005 to September 2005 The quantity of concrete placed at the bridge decks ranged from approximately 120 to 400 cubic yards The fresh concrete tests performed were slump (AASHTO T 119), unit weight (AASHTO T 121), air content (AASHTO T 152), and concrete temperature (AASHTO T 309)

For the first three decks, the research team performed all the fresh concrete tests

at three different locations (beginning, middle, and end regions) on the bridge deck and cast 4” x 8” cylinders for compressive strength tests at those same locations At the middle sampling location, eight 4” x 8” cylinders (for rapid chloride ion penetrability (RCIP) tests), four freeze-thaw specimens (3” x 3” x 16”), and four unrestrained shrinkage specimens (4” x 4” x 11.25”) were cast in addition to the compressive strength cylinders The last two decks were much smaller, and therefore the research team performed the fresh concrete tests at only two locations on the bridge decks Compressive strength cylinders were also cast at two locations on the smaller decks

The compressive strength cylinders (AASHTO T 22), freeze/thaw specimens (AASHTO T 161A), and unrestrained specimens (AASHTO T 160) cast at the first bridge deck were transported the morning after the deck placement, therefore complying with AASHTO T 23 However, for the four other bridge decks, the majority of the samples were transportated before the eight hour minimum time limit These samples were transported in the back of a full-size truck in containers that were placed on approximately three inches of soft foam to reduce vibration

After the initial 24 hours, the molds were removed and all specimens were air cured in an environmental chamber at 73°F and approximately 50 % relative humidity The research team measured compressive strength at 1, 7, 28, and 56 days of age Three cylinders were tested on each of these days The research team measured length change

at 1, 4, 7, 28, 56, and 112 days of age The research team tested RCIP at 28 and 90 days

of age, and freezing and thawing testing began at 14 days of age

Manual Crack Mapping

The process used to assess bridge deck cracking was similar to that used by AHTD Research Section personnel in a previous bridge deck study 13 The first step in the mapping process is typically an initial examination of the entire bridge deck to locate the areas most affected by cracking This is then the survey area It is generally limited to

100 ft in length unless the additional length would better represent the distress level of the deck as a whole Traffic control was provided by two AHTD personnel using a simple flagged lane closure with traffic cones along the centerline to keep motorists out

of the survey area

The actual mapping of distress in the survey section began with team members laying out a 100 ft tape measure along the lane edge This provided longitudinal stationing for the map A 25 ft tape measure was used to measure transversely from the

lane edges Cracks were then visually located and documented The length and location

of the cracks were measured and recorded, along with orientation and approximate widths of the cracks, on a prepared form with a grid representing the survey section The widths of the cracks were measured with a crack comparator card

AHTD Specifications for Class S(AE) Concrete

Concrete used in bridge decks in Arkansas are classified as Class S(AE) concrete (AE for air entrained) For Class S(AE) concrete, AHTD requires a minimum 28-day compressive strength of 4000 psi, a slump of 1 to 4 in., and an air content of 6 ± 2 percent AHTD also requires Class S(AE) concrete mixtures have a maximum w/cm of 0.44, a minimum total cementitious material content of 611 lb/ yd3, and a coarse aggregate meeting either the AHTD Standard Gradation or the AASHTO M43 #57 Gradation

AHTD allows the use of fly ash and slag cement in bridge decks Fly ash can either be Class C or F, with no mixing of the two The maximum fly ash replacement rate

is 20% by weight, and the maximum slag replacement rate is 25% by weight If both materials are used, the maximum replacement rate is 20%, by weight, for both materials

AHTD specifications allow the use of several different materials for concrete curing Burlap-polyethylene sheeting, polyethylene sheeting, copolymer/synthetic blanket, membrane curing compounds, and other materials that meet AASHTO M 171 are allowed AHTD specifications require that the bridge deck have curing compound applied be covered immediately after finishing, be covered using mats or blankets as a final cure and that it remains covered for at least 7 days During these 7 days, the curing materials must be kept continuously wet (except for membrane curing)

802.17(b) Application The exposed concrete, immediately after finishing, shall be

covered with one of the curing materials listed above and shall be kept continuously and thoroughly wet for a period of not less than 7 days after the concrete is placed Membrane curing does not require the application of additional moisture, except as required for bridge roadway surfaces

All Class B concrete shall be cured by free moisture Water curing shall be provided for all exposed surfaces for a period of 14 days

Membrane curing compound shall not be used on surfaces requiring a Class 2 finish

Clear membrane curing compound shall be used as an interim cure for concrete bridge roadway surfaces and shall be applied immediately after final finishing Final curing of bridge decks shall be by mats or blankets and shall be begun immediately after completing the surface test specified in Subsection 802.20(c) The mats or blankets shall

be kept continuously and thoroughly wet for a period of 7 days after the concrete is placed

AHTD allows the contractors to the option of continuous pours for bridge decks.

If the contractors choose this option, the concrete must remain plastic during the entire length of the pour Rather than casting the negative positive moment regions of the bridge deck first then followed by the positive negative moment regions, most contractors are choosing continuous casting or pours to spend speed up the construction process In this research program, all decks sampled were continuous pours

Concrete Mixture Proportions

The concrete mixture proportions for the five bridge decks are shown below in Table 1

As previously stated, AHTD requires a maximum w/cm of 0.44, a total cementitious material content of 611 lb/yd3, and an air content of 6±2% for bridge deck concrete All contractors chose to use the least amount of cementitous material required (611 lb/yd3) and three contractors chose to use fly ash (at replacement rates ranging from 15 to 20%) Three of the five concrete mixtures had the maximum w/cm of 0.44, and the lowest w/cm used was 0.41 A high range water reducer (HRWR) was used in the third bridge deck, which had the lowest w/cm The coarse aggregate content was different for all but two of the decks AHTD does not specify a coarse aggregate content Fly ash was the only SCM used in the bridge decks

TABLE 1 Concrete Mixture Proportions

Coarse Aggregate Type Limestone Limestone River Gr Limestone River Gr

BRIDGE SPECIFICS

As previously stated the research team sampled concrete from five bridge decks from June 2005 to September 2005 In addition to concrete properties, the researchers also documented the curing procedures and measured the cracking in each bridge deck Each

bridge deck is discussed in more detail in the following paragraphs The air temperature, relative humidity, and curing procedures are summarized for all decks and shown in Table 2

Bridge Deck 1

The first bridge deck visited is an interstate overpass The bridge deck was cast in the middle of June, and concrete placement began at 5:45 AM The bridge is 272 ft long and

43 ft wide and is a 2 span plate girder bridge with spans of 149 ft and 123 ft The total quantity of concrete used was 331 yards and the deck was a continuous pour

Like most bridge decks visited, the concrete was pumped up to the deck One construction worker with a commercial pressure washer fogged the concrete in the area

of placement The concrete was then screeded, floated with a pan attached to the finishing machine, and then manually tined with a rake Finally, a curing compound was applied and then the deck was covered with a plastic/cotton matcotton mat, burlap, and plastic sheeting

Bridge Deck 2

The second bridge deck visited was an interstate bridge overpass The deck was cast in the middle of July and concrete placement began at 9:00 PM The overpass bridge was built using staged construction and the portion cast on this date is 330 ft long and 32 ft 6

in wide and is part of a 4 span curved plate girder unit The total quantity of concrete used was 330 yards and the deck was a continuous pour

The concrete was pumped up to the deck One construction worker fogged the concrete at the surface near the finishing machine (prior to floating) The concrete was screeded and pan floated which was attached to the finishing machine The concrete was then bull floated with a 10 ft rounded float, and then manually tined with a rake The concrete was then sprayed with a curing compound and later covered with polyburlap for final cure

Bridge Deck 3

The third deck was a large city bridge that spanned a river The placement consisted of

400 yards of concrete and was a continuous pour The deck was cast in late August at 3:15 AM The plate girder bridge spans 367 ft with spans of 113, 141, and 113 ft The bridge deck is 43 ft wide

The concrete was pumped, screeded with the finishing machine, floated with a pan attached to the finishing machine, bull floated, and then tined with a finned float Like the previous decks, one construction worker fogged the concrete near the finishing machine using a pressure washer The concrete was then sprayed with curing compound and later covered with polyburlap

Bridge Deck 4

The fourth bridge deck was a state highway bridge that spanned a drainage ditch The bridge was placed in early September The three three-span bridge was a steel girder, w-section with spans of 38, 48, and 38 ft The bridge deck was 33 ft wide The placement consisted of a 117 yard continuous pour

The concrete was pumped, screeded with the finishing machine, floated with a pan and dragged with burlap that were both attached to the finishing machine It was then tined with a rake, sprayed with curing compound, and later covered with polyburlap

Bridge Deck 5

The final bridge deck is a US highway spanning a small creek The bridge deck was placed in late September The deck was a 171 yard continuous pour placement The concrete was pumped, screeded with the finishing machine, floated with a pan attached to the finishing machine The deck was bull floated with a 10 ft rounded float, and then manually dragged with burlap It was then tined with a rake, sprayed with curing compound, and later covered with polyburlap

TABLE 2 Summary of Observations for All Decks 1

Bridge

Deck

Time of

Placement

Size of Placement (yd3)

Air Temp

Range (°F)

Ave

R H (%)

Time to Curing Compound Application

Time to Final Cure1

Amount of Cracking ft/ft2

5 10:40 AM7:05 AM- 171 71-96 53 2.5 hr 5.00 hr 0.051

1 Times are from the end of the placement

RESULTS AND DISCUSSION

Crack Mapping

After the bridge decks were sampled, each deck was revisited to assess cracking For Bridge Deck 1, cracks were mapped on 4/5/06 after the bridge was open to traffic and

after the contractor had sealed larger cracks at some time prior to opening The researchers attempted to map all cracks in 12 by 100 ft section of south bound lane, but after measuring 40 ft of the 12 ft wide section, cracking became too small and random

to effectively map A 10 ft by 12 ft sub-area was measured as a representative sample From visual estimation, the density was approximately the same as the representative sample for the remainder of original 100 foot section, although it lessened some in last 15

ft The cracks ranged from 4 in to 48 ft in length and from less than 0.005 in to 0.024

in in width The cracks were a mix of transverse and longitudinal cracks with diagonal connecting cracks Long lines of cracking in the wheel path of the lanes were observed Also, cracks were concentrated over the center support (near the middle sampling location) of the deck This could possibly be due to a combination of vibrations from traffic passing under the bridge (which were noticeable) and low compressive strengths (at least up to 7 days) at this section

Bridge Deck 2 was revisited on 8/1/05 The visible cracks were measured for the whole pour The cracks ranged from 3 in to 17 ft in length and 0.002 to 0.016 in in width The cracks were mostly transverse and fairly heavy in the positive moment section The flexural cracks were located mainly near the piers and the plastic shrinkage cracks were located near the low gutter (the downhill side of the deck) Large amounts of paste were brought down to this side during construction using a highway screed High amounts of paste might have contributed to increased shrinkage in that area

Bridge Deck 3 was revisited on 1/27/06 The research team measured the cracking in a 12 ft by 100 ft section of the west bound lane The cracks ranged from 3

ft to 12 ft in length and were less than 0.007 in wide The cracks were almost exclusively transverse cracks that started and stopped at similar points in the cross section (near beam lines)

Bridge Deck 4 was revisited on 2/9/06 The research team measured cracking in a

12 ft by 100 ft section of the deck There was very little cracking in the deck The cracks ranged from 6 in to 4 ft in length and 0.002 to 0.010 in in width

Bridge Deck 5 was mapped on 2/10/06 The cracks were measured over a 12 ft

by 65 ft section The cracks were 2 ft to 5 ft in length and were 0.002 to 0.007 in wide There were some cracks that were 2 to 5 ft long and were at 45◦ angles to the intermediate bents

Fresh Concrete Data

As stated in the Testing Program, the fresh concrete properties were measured in two or three random locations (determined by AHTD) for each bridge deck If the bridge deck was large enough, the sampling locations were typically at the beginning, middle, and ends of the bridge deck The results from all the fresh concrete tests, the amount of cracking, and the AHTD specifications for each property are shown in Table 3

From Table 3, one can see that the four of the five bridge decks had slumps that exceeded AHTD specifications in at least one location Bridge Deck 1 was the only deck where all slumps fell within the 1 to 4 inch specification For the air content, three of the five bridge decks had measured air contents that did not meet AHTD specifications Only two bridge decks had fresh concrete temperatures that were greater than that allowed by AHTD

The final fresh concrete properties shown in Table 3 are the calculated and measured unit weights The calculated unit weights are based of off the concrete mixture proportion used by the concrete supplier and assuming a fresh concrete air content of 6% The differences between calculated and measured unit weights ranged from a low of 1 lb/ft3 to a high of 9 lb/ft3 These differences between calculated and measured unit weights could be attributed to the addition of extra mixing water or to higher or lower than expected air contents

In an attempt to determine of there were any relationships between the fresh concrete properties and crack density, the average slump, air content, measured unit weights, differences between measured and calculated unit weights, and concrete temperature were plotted versus the crack density Each bridge deck was ranked by each concrete property and assigned a ranking For example, Bridge Deck 1 had an average slump of 3.17 inches which was the lowest average slump of the five decks, and therefore

it received a ranking of “1” Likewise, Bridge Deck 4 had the greatest average slump (7.125 inches) and received a ranking of “5” Shown in the Figure 1 are the rankings for each fresh concrete property and crack density The graph shows that Bridge Deck 1, which had the highest crack density of 0.315 ft/ft2, did not have the greatest value for any

of the fresh concrete properties Bridge Deck 1 had the second highest air content, third highest concrete temperature, fourth highest unit weight, and was ranked last in unit weight difference and slump For the concrete properties measured and bridge decks samples, there was no correlation between fresh concrete properties and crack density

TABLE 3 Fresh Concrete Properties

Bridge Deck Slump(in)

Air Content (%)

Calculated Unit Wt

(lb/ft3)

Measured Unit Wt

(lb/ft3)

Concrete Temperature (°F)

Amount of Cracking ft/ft2 1

140

0.315

2

140

0.012

3

140

0.1125

AHTD

FIGURE 1 Fresh concrete properties and crack density.

Hardened Concrete Properties

The results from the compressive strength tests are shown in Table 4 Three cylinders for compressive strength testing were cast from either two or three random locations (as determined by AHTD) in each bridge deck The amount of cracking is also shown for each bridge deck in Table 4 For all decks, the contractors opted to pour each deck continuous, which by AHTD specifications, requires that all the concrete remain in a plastic state until concrete placement is finished Because of this reason, a set retarder was used in all decks

The first bridge deck that was visited (Bridge Deck 1) had the lowest one day strengths The first and last sampling location had a one day compressive strength of approximately 300 psi while the middle sampling location had a one day compressive strength of 60 psi At two days of age, cylinders that were sampled from the first and middle locations of the bridge deck were tested These tests showed that the first location had gained over 2000 psi in 24 hours, but the middle section was still much lower (a compressive strength of 130 psi) By 28 days and 56 days of age, the middle section had reached similar strengths as the first and last sections of the bridge However, the research team did observe several transverse cracks in the center section of the Bridge Deck 1 These cracks could be the result of the low compressive strengths of the middle location and the corresponding higher compressive strengths of the surrounding regions, but one cannot be certain due to the limited number of sampling locations

The only other bridge deck to have large variations in compressive strength was Bridge Deck 4 This deck was a smaller pour and due to time constraints only a limited