Non-Nuclear Methods for Compaction Control of Unbound Materials (2014)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

25 chapter three NON-NUCLEAR METHODS FOR DENSITY MEASUREMENTS OF UNBOUND MATERIALS INTRODUCTION Until the 1920s compaction of unbound materials was per- formed largely on a trial-and-error basis. Stanton (1928) was the first to use soil compaction tests to determine optimum moisture content and maximum dry density. Proctor (1933, 1945, 1948) extended this work and studied the effect of soil compaction on shear strength and permeability. He also significantly contributed to the development of the standard laboratory compaction test, commonly known as the Proctor test. Despite the many advances in compaction technologies since then, the Proctor test remains an important component in quality control compaction procedures of unbound materials in the United States. This chapter provides a review of the commonly used density-based compaction control methods. It also describes the non-nuclear density devices that have been evaluated as alternatives to the nuclear density gauge (NDG). Furthermore, it discusses the principles of operation and main advantages and limitations of those devices, thus synthesizing what has been reported in past studies. In the fol- lowing sections, photographs of non-nuclear density devices from certain manufacturers are provided for demonstration purposes only. This should not be construed as endorsements by this synthesis study of these devices. CURRENT DENSITY-BASED COMPACTION CONTROL METHODS The current compaction quality control of unbound materials involves determining the field dry density and moisture content of compacted lifts and comparing them to target density and moisture content values, which usually are determined using laboratory-specified tests performed on the same material used in the field. The ratio between the field density and the target laboratory density value is referred to as relative compaction. State DOTs typically require achieving a minimum relative compaction that varies between 90% and 100% for acceptance of compacted layers of unbound materials. It can be noted that the use of relative compaction requires that the material tested in the laboratory possess gradation and specific gravity simi- lar to that in the field. In addition, the field and laboratory compactive effort imparted to the material must be similar. According to Marek and Jones (1974), if the material type or imparted energy in the field differs significantly from the ref- erence material or compactive effort, the computed relative compaction will not be meaningful and valid. The following sections discuss the different methods that have been used to determine the target field density values for various unbound materials. Impact Compaction Laboratory Methods The impact compaction laboratory methods are the ones most commonly used to establish the compaction characteristics of unbound materials. These methods involve compacting a sample of the material to be used in the field in a standard mold using a drop hammer. The compaction energy can be varied by changing the number of hammer blows per layer as well as the weight and drop height of the hammer. Based on the compaction energy applied during the test, the proce- dure is either a “standard” or “modified” compaction proce- dure. The modified compaction test uses a compactive effort approximately 4.5 times greater than that of the “standard” test and was developed by the U.S. Army Corps of Engineers to better represent the compaction effort required for airfields to support heavy aircraft (Holtz et al. 2010). ASTM and AASHTO have standard specifications for both standard and modified compaction effort tests. The standard compaction procedure is documented as ASTM D698 or AASHTO T99, whereas the modified compaction procedure is specified as ASTM D1557 or AASHTO T180. Table 3 presents a com- parison between the ASTM and AASHTO specified compac- tion methods. Several state DOTs have developed and used modified versions of the original ASTM and AASHTO spec- ifications. Although these modified versions are somewhat different from the ASTM and AASHTO standards, the basic procedures and principles are the same (Tutumluer 2013). As shown in Table 3, the standard and modified compaction tests differ in terms of the hammer weight and drop height, as well as in the number of layers of unbound materials in the mold. Although the ASTM standards have three meth- ods (Methods A–C) for different mold sizes and maximum particle size of the material to be compacted, the AASHTO standards have four methods (Methods A–D). The selection of the appropriate method in both cases depends on the gra- dation of the tested material. The “standard” compaction method was originally devel- oped based on a study done by Proctor (1933). In this study, Proctor (1933) performed “plasticity-needle” penetration resis- tance tests to determine the indicated saturated penetration resistance (ISPR) for several compacted earth-fill dam sec- tions. Based on laboratory compaction tests, Proctor (1933)

26 Test Method Standard Effort Modified Effort ASTM (D698) AASHTO (T99) ASTM (D1557) AASHTO (T180) Mold diameter A 4 in. (102 mm) 4 in. (102 mm) B 4 in. (102 mm) 6 in. (152 mm) 4 in. (102 mm) 6 in. (152 mm) C 6 in. (152 mm) 4 in. (102 mm) 6 in. (152 mm) 4 in. (102 mm) D N/A 6 in. (152 mm) N/A 6 in. (152 mm) Mold volume A 0.0333 ft3 (944 cm3) 0.0333 ft3(944 cm3) B 0.0333 ft 3 (944 cm3) 0.075 ft3 (2,124 cm3) 0.0333 ft3 (944 cm3) 0.075 ft3 (2124 cm3) C 0.075 ft 3 (2,124 cm3) 0.0333 ft3 (944 cm3) 0.075 ft3 (2124 cm3) 0.0333 ft3 (944 cm3) D N/A 0.075 ft 3 (2,124 cm3) N/A 0.075 ft3 (2124 cm3) Rammer weight A, B, C 5.5 lbf (24.4 N) 10 lbf (44.5 N) D N/A 5.5 lbf (24.4 N) N/A 10 lbf (44.5 N) Height of drop A, B, C 12 in. (305 mm) 12 in. (305 mm) 18 in. (457 mm) 18 in. (457 mm) D N/A 12 in. (305 mm) N/A 18 in. (457 mm) Layers A, B, C 3 5 D N/A 3 N/A 5 Blows per layer A 25 25 B 25 56 25 56 C 56 25 56 25 D N/A 56 N/A 56 Compactive effort A, B, C 12,400 ft-lbf/ft3 (600 kN-m/m3) 56,000 ft-lbf/ft3 (2700 kN-m/m3) D N/A 12,400 ft-lbf/ft 3 (600 kN-m/m3) N/A 56,000 ft-lbf/ft3 (2700 kN-m/m3) Material A Passing No. 4 (4.75-mm) sieve Passing No. 4 (4.75-mm) sieve Passing No. 4 (4.75-mm) sieve Passing No. 4 (4.75-mm) sieve B Passing 3/8 in. (9.5-mm) sieve Passing No. 4 (4.75-mm) sieve Passing 3/8 in. (9.5-mm) sieve Passing No. 4 (4.75-mm) sieve C Passing ¾-in. (19-mm) sieve Passing ¾-in. (19-mm) sieve Passing ¾-in. (19-mm) sieve Passing ¾-in. (19-mm) sieve D N/A Passing ¾-in. (19-mm) sieve N/A Passing ¾-in. (19-mm) sieve Use A ≤25% by mass retained on No. 4 sieve ≤40% by mass retained on No. 4 sieve ≤25 by mass retained on No. 4 sieve ≤40% by mass retained on No. 4 sieve B ≤25 by weight retained on 9.5-mm sieve ≤40% by mass retained on No. 4 sieve ≤25% by mass retained on 9.5-mm sieve ≤40% by mass retained on No. 4 sieve C ≤30% by weight retained on 19-mm sieve ≤30% by weight retained on 9.5-mm sieve ≤30% by weight retained on 19-mm sieve ≤30% by weight retained on 9.5- mm sieve D N/A ≤30% by weight retained on 19-mm sieve N/A ≤30% by weight retained on 19- mm sieve N/A = not applicable. TABLE 3 SUMMARY OF ASTM AND AASHTO STANDARDS FOR IMPACT COMPACTION METHODS determined the compactive effort required to duplicate a field ISPR value of 300 psi. He suggested that soil samples be compacted in 1/20-ft containers using “firm blows” of a 5.5-lb tamper. However, as a result of a printing error, the “firm blows” were interpreted as “free-falls” of the tamper. This has led some organizations to assume that instead of striking a minimum length blow of 12 in., the tamper should be dropped a distance of 12 in. in free fall. Proctor pointed out this mistake in several of his published papers (Proctor 1945, 1948). He stated that the objective of compaction in earth fills was to achieve a penetration resistance value of 300 psi; the 12-in. blow was required to ensure accurate determina- tion of this value and was never intended as a “standard” or “optimum” (Proctor 1945). Despite Proctor’s strong recom- mendation against this, the use of soil dry density, instead of a strength-based penetration test, as the standard of soil compaction has been adopted by most organizations. In addition, the “standard” compaction method originally was intended for fine-grained soils. However, today it is

27 common practice to use impact compaction reference testing of other soil types for which it was not intended. Tutumluer (2013) notes that impact compaction reference testing may not adequately represent compaction characteristics in the field for certain aggregate that have low fines content. Both the ASTM and AASHTO standards can be conducted only on materials below a certain grain size, either 4.75 mm (0.19 in.) or 19.0 mm (¾ in.), depending on the method used. If the material to be tested includes particles larger than these sizes, corrections need to be applied to determine the unbound materials’ maximum dry density. The method typically used by state DOTs to perform this correction is AASHTO T224, “Correction for Coarse Particles in the Soil Compaction Test.” In this method, density is corrected by computing the weighted average of the density values of materials smaller and larger than the limiting particle size. Although density of the smaller material is determined using AASHTO T99 or T180, density of the larger material is based on knowledge of its bulk-specific gravity. This correction cannot be applied if the tested materials have more than 30% by mass of its particles larger than 19 mm (¾ in.). Several studies were conducted to evaluate the effect of the soil types and properties on the maximum dry density and optimum moisture content determined by impact compaction methods. Gregg and Woods (1960) reported typical ranges of maximum dry density and optimum moisture content values of different types of soils, which are summarized in Table 4. Several researchers have found that for fine-grained soils, good correlations exist between the maximum dry den- sity, the optimum moisture content, and the Atterberg limits (Woods 1938; Basheer 2001; Gurtug and Sridharan 2002; Omar et. al. 2003; Sridharan and Nagaraj 2005; Sivrikaya 2008; Sivrikaya et al. 2008; Kim et al. 2010). Based on tests results conducted on 102 soil samples in Indiana, Kim et al. (2010) proposed the equations shown in Table 5 to compute the compaction properties of fine-grained soils based on their plastic and liquid limits. One-Point Proctor Test The one-point Proctor method is an impact compaction test that was developed to determine the maximum density and optimum moisture content of unbound materials based on only a one-point measurement of density and moisture con- tent. In this method, a sample of the unbound material used in the field is obtained and compacted in a Proctor mold using a standard (AASHTO T99) or modified (AASHTO T180) effort. The dry density and moisture content is measured and plotted on predetermined compaction curves, referred to as the family of compaction curves, with similar shape and geometry for various types of tested soils. An example of such curves is provided in Figure 18. If the measured dry density and moisture content fall on one of the existing family of curves, the maximum dry density and optimum moisture AASHTO Classification Soil Description Anticipated Performance of Compacted Soil Typical Ranges of dmax OMC (%) pcf kN/m3 A-1-a, A-1-b Well-graded gravel/sand mixtures Good to excellent 115–142 18.1–22.3 7–15 A-2-4, A-2-5, A-2-6, A-2-7 Silty or clayey gravel and sand Fair to excellent 110–135 17.3–21.2 9–18 A-3 Fine sand Fair to good 100–115 15.7–18.1 9–15 A-4 Sandy silts and silts Poor to good 95–130 14.9–20.4 10–20 A-5 Elastic silts and clays Unsatisfactory 85–100 13.3–15.7 20–35 A-6 Silt-clay Poor to good 95–120 14.9–18.8 10–30 A-7-5 Elastic silty clay Unsatisfactory 85–100 13.3–15.7 20–35 A-7-6 Clay Poor to fair 90–115 14.1–18.1 15–30 After Gregg and Woods (1960). TABLE 4 TYPICAL RANGES OF MAXIMUM DRY UNIT WEIGHTS AND OPTIMUM MOISTURE CONTENTS USING STANDARD COMPACTION TESTS Parameters Considered in Developing the Relationship Relationship R 2 PL (%), dmax (pcf) PL = -48.024ln( dmax) + 245.82 0.66 LL (%), dmax (pcf) LL = -85.018ln( dmax) + 434.75 0.54 Source: Kim et al. (2010). TABLE 5 RELATIONSHIP BETWEEN gdmax, PLASTIC LIMIT AND LIQUID LIMIT

28 performing static compaction tests. However, the procedure that was followed in previous studies involved compacting about 2,500 g of moist unbound materials in a Proctor com- paction mold [typically a 102-mm (4-in.) mold was used] in one lift by applying static stresses using a hydraulic compres- sion device (Bell 1977; Zhang et al. 2005; White et al. 2007a). A load cell and two linear variable differential transformers (LVDTs) are typically utilized to measure the applied stresses and deformation, respectively. The obtained load versus defor- mation curve is used to determine the applied static compaction energy using Eq. 1. ) )( ( )( =Energy kN-m/m area of load versus deformation curve kN-m Volume of mold m (1)static 3 3 Aguirre (1964) compared the maximum dry density val- ues obtained using the static and impact compaction methods for 17 different soils. The results indicated that the moisture- density curves for both static and impact compaction tests were similar for coarse sands and gravels. However, the maxi- mum dry density determined using impact compaction for the plastic clay soils evaluated in that study were lower than those obtained in static compaction. Bell (1977) reported that at given moisture content, the static compaction required less compactive energy than both impact and kneading compaction methods to achieve a target density value. Zhang et al. (2005) and White et al. (2007a) reported that the static compaction energy required to obtain a given dry density value decreased with increasing moisture content. Bell (1977), Zhang et al. (2000), and White et al. (2007a) found that the moisture-density results were similar to impact compaction test results in their studies. content for that curve will represent those of the tested material. However, if the point (representing density and moisture content) falls in between two curves, a new curve that is parallel to and similar in shape to the nearest existing curve is drawn through it. The maximum dry density and optimum moisture content are then read from the new fitted curve. The one-point Proctor test is typically performed in accordance with the AASHTO T272-04 standard method. This test has been used by several state DOTs, including those of Indiana, Idaho, Iowa, Ohio, South Dakota, and Washington. Typically, each DOT develops its own family of compaction curves based on tests conducted on unbound materials encountered during construction. To evaluate the effectiveness of the one-point Proctor test, Wermers (1963) performed 861 compaction tests and com- pared the results from the one-point Proctor test with the standard Proctor test. He reported that the optimum moisture content found from the one-point Proctor test was on average 0.19% higher than that obtained by the standard Proctor test. Wermers also reported that 92% of the maximum dry density values obtained from the one-point Proctor test were within 4.0 pcf of those obtained from the standard Proctor test. Static Compaction Laboratory Method Porter (1930) introduced the static compaction method, in which he used a static pressure of about 13,800 kPa (2,000 psi) to compact granular soil samples of 152 mm (6 in.) in diame- ter. However, since that time, this method has not been widely used because the application of static pressure was not found to be effective in compacting granular materials (Rodriguez et al. 1988). Thus, currently there is no standard procedure for FIGURE 18 An example of the one-point Proctor method (Idaho DOT).

29 a controlled normal force to both the top and bottom of the sample at a constant gyration rate. The applied normal force was supplemented with a kneading action or gyratory motion at an angle (gyration angle) to compact the material. Currently, there are no standard values available on gyratory compaction parameters, such as pressure, angle, number of gyrations, or gyration rate. Different values of gyratory compaction param- eters were used in previous studies. A summary of these values is presented in Table 6. Smith (2000) reported a good correlation between labora- tory and field densities for well-graded crushed stone. For fine sand, Ping et al. (2003) found that the optimum moisture and maximum densities achieved in the field were closer to gyratory compaction results than impact and vibratory com- paction. In addition, they concluded that the gyratory com- paction method could be used to better simulate compaction conditions in the field. Similar conclusions were reached by Kim and Labuz (2006) and White et al. (2007a) for different types of granular and cohesive soils. Main Limitations of Laboratory Compaction Methods The main limitation of using laboratory compaction methods to select the target field density is that the volume of material used in those tests is very small compared with the total volume com- pacted in the field (Kim et al. 2010). If the compacted materials are highly variable, these methods will yield ambiguous results unless field corrections are made frequently. Another issue arises from the presence of a significant amount of gravel and cobbles in earth fill (Holtz et al. 2010). Although the use of laboratory impact reference testing has been extended beyond fine-grained soils, the AASHTO T99 and AASHTO T180 and their corresponding ASTM standards specify limits on the allowable amount of oversized particles in the tested material. For example, both test procedures are limited to unbound materials that have 30% or less by mass of particles with sizes greater than 19 mm (¾ in.). Because of the problems associ- ated with laboratory compaction methods, control strips (or test fill/strip) have been used by some DOTs to determine field target density value. Control Strip or “Test Strip” Method A control or test strip is a section that is compacted before and during fill-placement operations to determine the maximum target density value and the roller type, pattern, and number of Vibratory Compaction Laboratory Method In this method, the maximum dry density of unbound materials is determined by applying vibratory compaction forces. This can be done by using a vibratory hammer or a vertically vibrat- ing table. For tests conducted using a vertically vibrating table, there are three main variables that control the compaction energy imparted into the tested material, namely the vibration amplitude and frequency and the weight of the surcharge used. The compaction energy can be adjusted by changing any of those variables based on Eq. 2. Currently, there is an ASTM standard (ASTM D4253) for performing vibratory compaction tests, but no such specification is provided by the AASHTO. W f A t Vvib = × × ×Energy (2) where Energyvib = vibratory compaction energy (kN-m/m3), W = weight of surcharge (kN), f = frequency of vibration (Hz), A = amplitude (m), t = time (s), and V = volume of mold (m3). Lambe (1951) reported that granular free-draining soils do not respond to variations in compacting moisture content and impact energy as cohesive soils because of negligible lubrication; therefore, vibratory compaction should be used. Many researchers reported that vibratory compaction methods produce consistently higher maximum densities for granular materials than does the impact compaction method and also better replicates field densities (Burmister 1948; Felt 1958; Pettibone and Hardin 1964). Although the vibratory compaction method was developed originally for granular soils, some studies indicated that it can be effective in cohesive soils if compacted at low frequencies (Converse 1956; Lewis 1961). According to a recent survey conducted by Tutumluer (2013), only two state DOTs (Kansas and Alabama) reported the use of vibratory com- paction methods for unbound aggregate materials. Gyratory Compaction Laboratory Method The U.S. Army Corps of Engineers introduced the gyratory compaction test procedure for soils based on extensive testing on silty sand material (Coyle and West 1956; McRae 1965). The test procedure consisted of placing a thoroughly mixed, loose, moist sample in a cylindrical mold and then applying Study Vertical Stress (kPa) Gyration Angle No. of Gyrations Soil Type Smith (2000) 1,380 1.0 30–40 Well-graded crushed stone Ping et al. (2003) 2,000 1.25 90 Fine sand Kim and Labuz (2006) 6,000 1.25 50 Recycled granular material White et al. (2007a) 6,000 1.25 50 Granular and cohesive soil TABLE 6 GYRATORY COMPACTION PARAMETERS USED IN PAST STUDIES

30 Solid Volume Density Method In this method, the target field density is selected as a per- centage of solid volume density. The solid volume density represents the density of soil solids in a voidless matrix, which can be obtained by multiplying the specific gravity of the aggregate with the unit weight of water. The constructed layer densities in this method are then expressed as a per- centage of the solid volume density, referred to as relative solid density (RSD) or solid relative density (SRD). In this method, the relationship between the achieved densities in the field and solid volume density of the compacted material should be known. Kleyn (2012) reported the application of the solid volume density method in the construction con- trol of G1 base in South Africa. He found that 88% of SRD was equivalent to about 106% of the maximum dry density obtained using the modified compaction method. In addi- tion, he indicated that there are a number of conditions that need to be satisfied before using this method. Kleyn (2012) stated that the aggregates have to be very resistant to gen- eral construction impacts and free from contamination or deleterious materials. Only fresh, unweathered, and sound passes needed to achieve this density of a particular material. The strip is compacted at a moisture content close to the optimum using the compaction equipment to be utilized by the contractor. Field density and moisture measurements are obtained at three or more randomly selected locations after each pass until no significant increase in density is observed. The average final density of the material from the control strip is defined as the maximum target density for that particular material. Usually agencies specify that lifts must be compacted to a certain percentage of this maxi- mum density. Test sections also are typically constructed every 1,000 to 3,000 m3 (1,500 to 4,000 yd3) or where the compacted material changes significantly. It can be noted that the compaction of a control strip should be correlated to previously established compaction results for success- ful implementation of the control strip method (Tutumluer 2013). Based on the review of state DOT construction spec- ifications and manuals, it was found that several states have specifications for using control strips in their compaction control procedures for unbound materials. Table 7 presents a summary of state DOT control strip specifications for unbound materials. Agency Length (minimum) Width (minimum) Target Value Alaska 300 ft 12 ft 95% of the maximum control strip density measured using NDG District of Columbia 100 ft At least one lane wide Not specified Maryland 100 ft At least one lane wide Not specified Kentucky 500 ft Full lane width Five tests must be at least 98% of the target density with no individual measurement less than 95% of the target density Mississippi 500 ft 12 ft Same rolling pattern and number of passes used in test strip (only for aggregate drainage layer) North Carolina 300 ft Full lane width Not specified New Hampshire 100 ft Full lane width 95% of maximum control strip density measured using NDG Virginia 300 ft Full lane width Average 98%; individual minimum 95% of control strip density New Hampshire 100 ft Full lane width Not specified Alabama 500 ft Not specified 100% Indiana 100 ft Lane width Not specified Minnesota 300 ft 32 ft 90% of IC target value Michigan 600 ft Not specified 95% of the maximum unit weight; used for open-graded drainage course base only New Jersey 400 square yards Q = South Dakota 500 ft Not specified 95% of target density West Virginia 100 ft Full width Not specified ≥ 0.36 TABLE 7 SUMMARY OF STATE DOTs’ CONTROL STRIP SPECIFICATIONS FOR UNBOUND MATERIALS

31 moisture content and density measurements. The main draw- back with the gauge is that it uses radioactive materials that necessitate strict compliance with regulatory requirements for handling, storage, maintenance, transport, and monitoring. Furthermore, NDG measurements may be affected by the chemical composition of the soil tested (ASTM D6938). Spe- cifically, moisture content measurements are affected when hydrogen atoms exist in the chemical composition of the soil and other recycled pavement materials commonly used today. Because of these issues, highway agencies, universi- ties, and equipment manufacturers have developed several new methods and devices to replace the NDG. Some stud- ies were performed to evaluate these devices. The follow- ing sections provide a review of the different non-nuclear density methods that have been used or evaluated by state DOTs during the past decade. In addition, a summary of the main findings of previous studies on these devices is presented. Moisture Density Indicator The moisture density indicator (MDI) uses time domain reflectometry (TDR) to measure the dry unit weight and moisture content of soils. It consists of four metal spike probes that are encased in a probe head, which is connected by a coaxial cable to a TDR pulse generator. The genera- tor is attached to a personal digital assistant (PDA) with proprietary software. The MDI works by sending an elec- tromagnetic wave pulse through the four probes that are driven into the soil in the formation shown in Figure 20 a to imitate a coaxial cable configuration. The center spike acts as the central conductor in the coaxial cable, the three outer probes serve as the shield conductor, and the in situ soil acts as the insulator. Typically the spikes have diameters of 0.75 in. and variable lengths of 4, 6, and 8 in. (Rathje et al. 2006; Jackson 2007). The signal reflected back through probes is recorded and analyzed by the PDA using proprietary software to determine electrical proper- ties of the tested soil. This is used to determine the dry density and moisture content of the tested soil. This test is performed according to ASTM D6780. The MDI costs about $6,000. rock could be used in construction of the base layer in this procedure. Currently, the solid volume density method has not yet been used in the United States or Canada. MEASURING IN-PLACE DENSITY OF CONSTRUCTED UNBOUND MATERIALS Most state DOTs determine the in situ dry density and mois- ture content of compacted unbound materials using the nuclear density gauge (NDG) (shown in Figure 19). The device was introduced in the early 1970s and gained popularity after an industrywide calibration standard was developed for it (Kim et al. 2010). The NDG works by emitting gamma radiation into the material to be tested and detecting the reflected rays to determine its wet density. Denser materials contain more electrons with which the photons of the gamma radiation interact; therefore, they reflect a lower number of photons. The number of detected photons is used to calculate the density of the tested material based on calibrated relation- ships. In addition to the density measurement, the NDG is capable of measuring the moisture content of soil. The high- speed neutrons emitted from the nuclear gauge source get retarded by the hydrogen atoms present in the moist material. The number of slow-speed neutrons detected by the gauge is used to determine the amount of hydrogen atoms present in the material, which is then used to compute its moisture content (ASTM D6938). Nuclear gauges can be operated in two modes: direct transmission mode and backscatter mode. Winter and Clarke (2002) reported that direct transmission mode yielded a more accurate density measurement than did backscatter mode. The main advantage of the nuclear gauge test over other conventional tests is that it is relatively fast to perform. In addition, it is accurate and repeatable when properly cali- brated. The NDG can be used to measure the density of asphalt concrete as well as layers of unbound materials. It also has the advantage of being able to vary depth of measurements using the direct transmission procedure. The NDG is also one of the few density tests that has the capability to provide both FIGURE 19 Nuclear density gauge (Troxler 2000).

32 where Vs = voltage source, Vf = long-term voltage level, C = constant related to probe configuration, Rs = internal resistance of pulse generator, do = outer conductor diameter, and di = inner conductor diameter. Equations 6 and 7 relate the Ka and ECb directly to the dry unit weight and moisture content of the soil (Durham 2005; Rathje et al. 2006). (6)1 2 pK a b wa w d( )( ) ρ ρ = + (7)1 2 pEC c d wb w d( )( ) ρ ρ = + where a, b = soil-specific calibration constants for Ka, c, d = soil-specific calibration constants for ECb, Two main electrical properties of tested soils are measured using the MDI: the dielectric constant, Ka, and bulk electrical conductivity, ECb. Ka is calculated using Eq. 3 based on the first and second reflections in the recorded TDR waveform, which represent when the electromagnetic wave enters the soil and when it is reflected from the end of the MDI soil spikes. In addition, the bulk electrical conductivity (ECb) is found by measuring the source voltage (Vs) and final voltage (Vf) of the TDR waveform using Eq. 3 (Rathje et al. 2006). (3)2K L La a p( )= where La = scaled horizontal distance between the first and sec- ond reflection points in the TDR waveform, and Lp = the length of the soil probes. EC C V Vb s f( )( )( )= −1 1 (4)p 2 ln (5)p p pC L R d dp s o i( )( ) ( )= pi (a) (b) (c) (d) FIGURE 20 MDI field testing: (a) probe configuration, (b) probes being driven into the ground, (c) coaxial head placed on top of the soil probes, and (d) field measurement being taken (Durham 2005).

33 Berney et al. (2013) reported that the installation of the four probes in the seating mold caused it to become wedged into the soil surface. This made the removal of the mold difficult without disturbing the probes, resulting in a loss of contact with the soil along the entire probe length, which led to a very low moisture content reading. This effect was exacerbated in soils with coarser grain sizes. Finally, some studies suggested that the operation of the MDI required an operator with at least moderate knowledge of the device’s overall capabilities. Synthesis of Past Research Studies New Jersey Jackson (2007) conducted a study for the New Jersey DOT to evaluate the effectiveness of the MDI as a tool for compaction control of dense-graded aggregate base lay- ers. Field testing was conducted on five sites that consisted of dense-graded, aggregate base layers as well as on some New Jersey DOT-designated porous fill materials. The results of this study indicated that both the nuclear gauge and the MDI recorded similar moisture contents. However, differences of as much as 12.53% were observed in the dry density mea- surements. In general, the dry density values measured by the MDI were less than those obtained using the nuclear gauge. The MDI appeared to be less sensitive to the changes in com- paction density measured at different locations at a given site as compared with the NDG. Jackson (2007) indicated that it was difficult to drive and remove the spikes into the aggregate base. A much larger hammer was needed, and it took more than 15 min per test to drive in the spikes. He suggested that spikes of a least 1 in. in diameter be used to enhance penetra- tion of the compacted aggregate base. Jackson (2007) also reported that the MDI data acquisition software froze several times during field and lab testing. Finally, from a job-site effi- ciency standpoint, researchers found that the transportation of the MDI device was time consuming and cumbersome. Vermont Brown (2007) reported the results of a study conducted to evaluate the performance of two non-nuclear density devices (EDG and MDI) and compare them with that of the NDG. As shown in Figure 21, the dry density values measured by the MDI had a very good correlation with those of the NDG. However, the moisture content mea- sured by those devices showed high variability, as shown in Figure 22. Brown (2007) indicated that the MDI was time consuming to set up and was not easily transported around the construction site. Its many loose parts required multiple trips to move the device from one spot to another. In addition, Brown (2007) found that MDI setup did not work well in coarse materials because the MDI uses four spikes that are driven through a template in a very concen- trated area (of usually about 8 in.). In addition, the spikes were prone to bending when used on aggregate materials or densely compacted subgrade material. Florida Sallam et al. (2004) presented the results of a study in which the TDR was used to measure moisture contents of A-3 (fine sands) and A-2-4 (sands and gravels with elastic silt fines) at different sites. Figure 23 shows the moisture content rw = unit weight of water, rd = dry unit weight of soil, and w = moisture content. The MDI has two operation modes: one step and two step. In one step mode, the bulk electrical conductivity (ECb) and dielectric constant (Ka) values are simultaneously measured for a given soil used to determine the dry density and moisture content. In two step mode, the MDI is used to measure the Ka of the soil in place and a soil sample excavated from the field and compacted in a mold. The density of the in situ soil is deter- mined using the density of the soil in the mold and the dielectric constants measured in the field and mold. Calibration of the MDI requires determining the constants in Eqs. 6 and 7 for the specified soil. This is achieved by measuring the Ka and ECb for several samples compacted in a Proctor mold at a range of known dry densities and mois- ture contents. The obtained data are used to develop plots of (Ka)1/2(rw /rd) and (ECb)1/2(rw /rd) versus moisture content, and a line is fit to the plotted data to determine calibration constants of tested soil. Repeatability and Accuracy In general, all studies indicated that the MDI is a repeatable device. The reported coefficient of variation of MDI mea- surements was in general less than 15% (Rathje et al. 2006). Previous studies showed that MDI moisture content measure- ments were accurate and very close to those obtained using the oven dry method. The MDI and NDG differed in their dry density measurements (Jackson 2007; Ooi et al. 2010). Ooi et al. (2010) questioned the MDI ability to yield reliable results because of a flaw in its equation formulation; they also suggested that its formulation should be reevaluated. Main Advantages and Limitations The MDI has some advantages over the NDG. First, the MDI does not require special licensing to operate (Rathje et al. 2006; Brown 2007). In addition, the MDI operator depen- dency does not contribute to significant variability in its mea- surements. The main disadvantage of the MDI is that it is time consuming (15 to 20 min per test). In addition, operating the MDI is more cumbersome than other methods, as many loose parts are required for operation. In addition, it may require excavation of soil from the construction site. Some researchers also indicated difficulty in driving and remov- ing the spikes into the aggregate base or stiff subgrade soils. The MDI also presents limitations on the types of soil it can test. For example, it cannot be used for highly plastic clays because of issues related to electrical conductivity in those clays (Yu and Drnevich 2004). Furthermore, the MDI can test soils with 30% or less, by mass, of its particles smaller than 4.75 mm (0.19 in.) and has a maximum particle size of 19 mm (¾ in.). MDI cannot be used to test frozen soils.

34 FIGURE 21 MDI dry density compared with NDG dry density (Brown 2007). FIGURE 22 MDI moisture content compared with NDG moisture content (Brown 2007). FIGURE 23 MDI moisture content compared with NDG moisture content (Sallam et al. 2004).

35 Texas Rathje et al. (2006) conducted field and laboratory testing programs to evaluate the MDI. The field component of the study included determining the density of clayey soils used in Texas for highway embankments or as road subgrade. In addition, the laboratory portion of the study included using the MDI and other devices to test laboratory- compacted samples of poorly graded sand. The results of this study indicated that for clayey soils, the MDI dry density measurements did not agree favorably with those obtained using the nuclear gauge or the rubber balloon. Rathje et al. (2006) also indicated that the MDI dry density values were higher than those of the NDG for high plasticity clays and lower for low plasticity clays. Furthermore, the MDI mois- ture contents were different than the values measured by the traditional oven drying method and the nuclear gauge. For the laboratory test samples of sand, the MDI showed good agreement with the microwave oven measurements of moisture content. However, it consistently reported the same dry density value for all samples, which did not agree with the rubber balloon measurements. Electrical Density Gauge The electrical density gauge (EDG) (Figure 25) uses high radio frequency waves to measure the density and moisture content of soils. It consists of four tapered 6-in. long spike probes, a hammer, a soil sensor and cables, template, temperature probe, battery charger, and hard case. The device works by transmit- ting high radio frequency waves through the four probes that are driven into the soil in a square formation. Four measure- ments are obtained at each test location after the probes are inserted. The EDG analyzes the transmitted radio frequency to determine the electrical dielectric properties of the tested data that were obtained in that study. The TDR recorded less scattered and closer moisture content values to the labora- tory oven method compared with the nuclear gauge. Thus, it yielded more repeatable and accurate measurements. In gen- eral, the TDR consistently underestimated the moisture con- tent, whereas the nuclear density method overestimated it. Runkles et al. (2006) conducted a study that evaluated the accuracy and repeatability of TDR’s one-step method (MDI one step mode) in measuring the dry density and moisture content of sandy soil typically encountered in construction in Florida. The results indicated that the TDR method showed lower variability in the moisture content measurements than did the NDG and could be more accurate if the calibration constants were properly selected. In addition, as shown in Figure 24, the TDR method had a better correlation with the oven dry moisture content measurements than did the nuclear gauge. However, Runkles et al. (2006) reported that the TDR recorded more variable and lower density measure- ments than did the nuclear gauge. Hawaii Ooi et al. (2010) reported the results of a study that compared the moisture content and dry density of recycled concrete aggregate (RCA) and reclaimed asphalt pavement (RAP) obtained using nuclear gauge and TDR methods with actual values by compacting these materi- als in 6-in. lifts and 3-ft diameter bins. The results indi- cated that the TDR moisture content measurements for RAP and the dry densities for RAP and RCA were reasonably accurate. However, the TDR underestimated the moisture content of the RCA material. The authors also reported that the TDR provided more accurate dry density measure- ments than did the NDG when the RCA material was tested 10 days after its compaction. FIGURE 24 MDI moisture content compared with NDG moisture content (Runkles et al. 2006).

36 ever, this was found to depend mainly on EDG calibration for the considered material (Jackson 2007; Von Quintus et al. 2008; Berney et al. 2013). In addition, Berney et al. (2013) indicated that the relative error in EDG moisture content measurements was approximately 4% when using the lab oven dry method measurements as a reference. On the contrary, Rathje et al. (2006) reported a less favorable agreement between EDG dry unit weight and water content measurements with traditional method measurements when testing poorly graded sand. Cho et al. (2011) found that aver- age error in the EDG dry density and moisture measurements was much higher than that of the NDG. Main Advantages and Limitations The main advantage of the EDG over the NDG is that it is safer and does not require special licensing to operate (Rathje et al. 2006; Brown 2007). However, the EDG calibration pro- cess was found to be complex and time consuming (Rathje et al. 2006; Berney et al. 2013). Previous studies indicated that the EDG was cumbersome and time consuming to oper- ate in the field because the switching of connectors to and from different probes takes some time. Furthermore, similar to the MDI, the device had many loose parts (template, darts, hammer, and electrodes), making it difficult to transport from one location to another at the site. The probes also could be difficult to drive into stiff soils. Rathje et al. (2006) also sug- gested that the EDG could not be used for high-plasticity clays and said it relies heavily on other density tests for its calibration. Synthesis of Past Research Studies Vermont The results of a study conducted by the Vermont DOT indicated that the EDG’s dry density measurement had a strong correlation with that of the NDG (shown in Figure 26) (Brown 2007). However, the EDG moisture con- tent measurement, presented in Figure 27, had a weak cor- relation with that of the NDG. The authors attributed this to the high variability of moisture contents within various soil types and soil depths. Texas As part of a study conducted by the Texas DOT, Rathje et al. (2006) evaluated the ability of the EDG to accurately measure the density and moisture content of lab-compacted samples of poorly graded sand. The EDG constantly measured moisture content values of 5% for all samples, whereas the values recorded by the micro- wave oven ranged between 2.9% and 3.4%. In addition, although the EDG dry density measurements were about 90 pcf for all soil samples, the rubber balloon measure- ments were 100 to 115 pcf. Rathje et al. (2006) concluded that there was no agreement between EDG density and moisture content measurements and those obtained using traditional methods. soil, which include resistance (Rs), capacitance (Cs), the quo- tient (Cs/Rs), and real impedance (Zs). Those properties are converted to dry density and moisture content measurements by using a soil-specific calibration model, which is devel- oped by taking EDG readings of the soil samples compacted in a laboratory mold at different moisture contents and den- sity combinations to determine dielectric constants for each combination. The EDG is conducted in accordance with an ASTM D7698 standard. However, a detailed description regard- ing the theoretical basis for the EDG is not currently available. The price for the EDG device is approximately $9,300 (plus $2,250 for the calibration verifier) (Humboldt Mfg. Co. 2012). The EDG has GPS capabilities as far as 3 meters away from the location of the probes (Humboldt Mfg. Co. 2012). Repeatability and Accuracy There is no consensus in the literature on the repeatability and accuracy of the EDG. As part of NCHRP 10-65, Von Quintus et al. (2008) reported that the EDG was a highly repeatable device and had a coefficient of variation (COV) of dry den- sity and moisture content of less than 1% and 5%, respec- tively. Berney et al. (2013) reached similar findings when testing various types of soils. However, Brown (2007) indi- cated that there was high variability in EDG moisture content measurements. In terms of accuracy, some studies showed that the EDG recorded dry density measurements that were similar to those obtained using traditional methods such as NDG (Brown 2007; Von Quintus et al. 2008; Berney et al. 2013). How- FIGURE 25 Electrical density gauge (Berney et al. 2013).

37 age difference was 1.71 pcf and 0.22%, respectively. Cho et al. (2011) attributed those results to the nuclear gauge data being corrected using the density and moisture cor- rection factors required by the Nebraska DOT standard test method for the NDG. In addition, they noted that the EDG had similar results to the NDG before the correction factors were applied. Based on the results of life-cycle cost analy- ses, Cho et al. (2011) indicated that despite the high initial cost of the EDG, it presented an economic advantage over the nuclear gauge when maintenance and operating costs were included. NCHRP Project 10-65 Von Quintus et al. (2008) reported the results of NCHRP 10-65, in which the repeatability and accuracy of the EDG were evaluated for selected unbound materials. Figure 30 compares the obtained EDG dry den- sity values to those measured with the traditional NDG. There are two different groups of data: one for fine-grained soils and the other for crushed aggregate base materials. In general, as the NDG dry density increased among differ- ent materials, the EDG density also increased. However, no apparent correlation exists between EDG and NDG dry density measurements. Nebraska Cho et al. (2011) reported the results of a study that evaluated the effectiveness of the EDG in measuring in situ moisture content and density. The study included conducting EDG and NDG tests on several soil types at several construction sites in Nebraska. The EDG and NDG measurements were also compared with the standard field dry weight unit measurement determined by taking a sam- ple representative of each measurement area either with a Shelby tube or other method for lab testing. Figure 28 shows the relationships between the EDG and the NDG dry density measurements and those obtained using the drive- cylinder method. In addition, Figure 29 presents the com- parisons of the moisture content measurements of the EDG and the NDG to those obtained using the standard oven dry test method. The NDG results had a better correlation than did the EDG results to the moisture content and dry density measurements that were obtained using the stan- dard methods (i.e., oven dry test and drive-cylinder meth- ods). In addition, the average difference between the EDG dry density and moisture content measurements and those obtained using the standard measurement methods were found to be 9.86 pcf and 1.66%, respectively; for the NDG dry density and moisture content measurements, the aver- FIGURE 26 EDG dry density compared with NDG dry density (Brown 2007). FIGURE 27 EDG moisture content compared with NDG moisture content (Brown 2007).

38 FIGURE 28 Dry densities measured with EDG and NDG (Cho et al. 2011). FIGURE 29 Moisture content measured with EDG and NDG (Cho et al. 2011). FIGURE 30 Dry densities measured with EDG and NDG (Von Quintus et al. 2008).

39 Soil Density Gauge The soil density gauge (SDG) is a self-contained unit that uses electromagnetic impedance spectroscopy (EIS) to mea- sure the density and moisture content of various unbound materials. As shown in Figure 32, the SDG measurement is done through a noncontacting sensor that consists of a central ring and an outer ring. The central ring generates and trans- mits a radio-frequency–range electromagnetic field into the soil. The response to that field is received by the outer ring and is used to measure the dielectric properties of the tested soil matrix. The SDG performs a calculation on the measured dielectric properties to determine the density and moisture content of the tested soil (TransTech Systems, Inc. 2008). The price of the latest SDG model (SDG 200) starts at $10,000. This model has an advanced GPS system that enables location and independent time logging. The SDG should be calibrated by compacting a sample of the soil of interest in a Proctor mold and testing it using SDG. The density and moisture content of the compacted soil sample are used as initial condition values for calibration (Gamache et al. 2009). The SDG test consists of taking five individual measurements in a counterclockwise “cloverleaf” pattern at the test site. The first location measured represents the place where density and moisture content is desired to be obtained, and the outline circles of the other four locations must be 1 to 2 in. away from the initial center circle measure- ment. The surface where the SDG is placed for testing must be flat and free of small stones or debris for the consistency of results (TransTech Systems, Inc. 2008). Accuracy and Repeatability In general, previous studies indicated that SDG density and moisture content were repeatable and very close to the NDG values once the linear offset correction was made to the U.S. Army Corps of Engineers Berney et al. (2013) reported a study conducted at the U.S. Army Corps of Engineers Research and Development Center in Vicksburg, Missis- sippi, to evaluate the effectiveness of various devices, includ- ing the EDG, in controlling the compaction of soils used for horizontal construction. The results indicated that the EDG was the second most effective electrical device for measuring the dry density of the various types of soils, hav- ing the most accurate precision but only average accuracy when compared with the NDG. Berney et al. (2013) found that the EDG performed better in fine-grained soil than in granular material. In another study, Berney et al. (2011) examined the repeat- ability and accuracy of EDG moisture content measurement. Figure 31 presents the data collected in this study. The EDG moisture content had a strong correlation with that of the laboratory oven method. In addition, both methods yielded very close values, as indicated by the slope of the line in Fig- ure 31. This suggests that the EDG had good accuracy and was as repeatable as the laboratory oven method. FIGURE 31 Moisture content of EDG and laboratory oven dry method (Berney et al. 2011). FIGURE 32 Soil density gauge, SDG200 (Pluta et al. 2008).

40 improved by accounting for the specific surface area of the material being tested. U.S. Army Corps of Engineers Berney et al. (2011) evaluated the performance of the SDG device in measuring moisture content of various types of soils used by the U.S. Army Corps of Engineers for horizon- tal construction. As shown in Figure 35, raw SDG moisture content measurements were highly scattered and had a poor relationship with those obtained using the laboratory oven method. Berney et al. (2011) used Eqs. 8 and 9 to correct the SDG measurement by applying a linear offset, which repre- sents the difference between a recorded SDG moisture con- tent measurement and that obtained using the oven method for the soil of interest. Figure 35 shows that the corrected SDG value had much better correlation with the laboratory oven method moisture content measurements. In addition, the corrected SDG values were also closer to those of the lab oven method than the NDG measurements. Based on that, Berney et al. (2011) recommended that the SDG measure- ments be corrected using Eqs. 8 and 9. Berney et al. (2013) conducted a study that assessed the performance of the SDG and other non-nuclear density devices. The results indicated that the SDG was the most effective device overall, possessing an optimal combination of accuracy and precision compared with the NDG. Berney et al. (2013) also found that the SDG had better performance in granular soils compared with fine-grained soil. This was attributed to the SDG manufacturer developing its platform using more granular soil types. Therefore, the researchers recommended that the SDG be tuned to capture the density variance in wetter, fine-grained soils. device’s results (Pluta et al. 2008; Berney et al. 2011). How- ever, Pluta et al. (2008) reported that soil gradation affected the frequency response of the SDG, which could reduce the repeatability of this device because of inconsistencies. Berney et al. (2013) also found an overall coefficient of variation of the corrected SDG to be 0.278 with regard to moisture content. Advantages and Limitations The SDG can provide accurate and repeatable moisture con- tent and density measurements when the proper corrections are applied and the gradation of the tested unbound materials is consistent with its calibration. It was suggested that opera- tors of the SDG have extensive knowledge of this device in order to apply the needed corrections (Berney et al. 2011). Synthesis of Past Research Studies Pluta et al. (2008) conducted SDG and NDG tests on vari- ous types of fine- and coarse-grained soil. Figure 33 pre- sents the obtained wet density measurements in that study. The soil type had a significant effect on the SDG density measurements. To address this issue, the authors applied a linear adjustment based on the specific surface area of the material being tested to correct the SDG soil model’s calcu- lation of wet density and moisture. Figure 34 shows the wet density values that were corrected by applying the surface area adjustment. In five of the six types of tested soils, the corrected SDG wet density measurements came much closer to those of the NDG. In addition, the average difference in wet density between the NDG and the SDG was reduced by 119%. Pluta et al. (2008) concluded that the accuracy and precision of the SDG density measurements could be FIGURE 33 SDG wet density compared with NDG wet density (Pluta et al. 2008).

41 MOISTURE MEASUREMENT Whether measuring density, modulus, or shear strength, moisture content remains a critical parameter in compac- tion quality control procedures of unbound materials. There- fore, it is essential to obtain rapid, reliable, and accurate moisture content measurements of compacted unbound materials in the field. There are several non-nuclear methods that have been used for measuring the in situ moisture content of unbound materials during construction. The fol- lowing sections discuss the main methods that have been . . (8)oven #1L O STD MCSDG= − . . (9)CorrSDG MC L OSDG= + where L.O. = linear offset, STDoven = laboratory oven moisture content (standard), MCSDG#1 = first SDG moisture content for a specific soil, SDGCorr = corrected moisture content for that soil, and MCSDG = SDG device reading for moisture content. FIGURE 34 Corrected SDG wet density compared with NDG wet density (Pluta et al. 2008). FIGURE 35 Corrected SDG and initial SDG moisture content compared with moisture content obtained using the laboratory oven method (Berney et al. 2011).

42 soil lumps, improper sealing of the vessel, insufficient time allotted for the chemical reaction, and the presence of vola- tile material in the tested material (Petersen et al. 2007). Cal- cium carbide is considered a hazardous material that requires special handling and environmental considerations. Sallam et al. (2004) indicate that one source of error for this method is the operator’s ability to perform the test correctly. Synthesis of Previous Studies Oman (2004) presented the results of a study in which the moisture content was obtained using the speedy moisture tester and oven dry methods for various types of granular soils in Minnesota. The results indicated that the methods provided comparable moisture content measurements. In addition, as shown in Figure 37, a strong correlation existed between the two methods. Alleman et al. (1996) found that the speedy moisture tester overestimated the moisture content by 1.25%. However, it was considered reliable once calibration was performed. George (2001) found that the results from the speedy moisture tester were comparable to results from the NDG. Runkles et al. (2006) reported the results of side-by-side tests performed to measure the moisture content of common construction soils in Florida using the ASTM TDR one-step method, nuclear gauge, oven dry, and speedy moisture methods. The results indicated that the speedy moisture method was slightly more variable than the ASTM TDR and nuclear methods. In addition, it had poor cor- relation with the oven dry measurements. The ASTM TDR and nuclear method measurements had a much stronger correlation with the oven dry measurements. Berney et al. (2011) reported that the results of a study to evaluate the effectiveness of various devices to measure the moisture content of soil for horizontal construction. The accu- racy and repeatability of the considered devices were compared with the standard laboratory oven soil moisture determination. The results of this study indicated that this device had the most effective repeatability; however, it had the worst accuracy among other moisture devices that were used. As shown in Figure 38, Berney et al. (2011) found that the speedy moisture device overestimated the moisture content for all soil types for different reasons. For coarse-grained soils, small sample size (only 20 g) was found to be the reason for the overestimation of moisture content. The sample size tends to contain only fine- grained material, which retains the most available moisture. In addition, for fine-grained soils with high moisture content, the speedy moisture tester required a multiplier to be applied to the charted conversion values. This multiplication increased the overall error in the moisture content results. Moisture Analyzer The moisture analyzer is a small drying device with a scale and an overhead ceramic heating element (Figure 39). To use evaluated by state DOTs other than those discussed in the preceding sections. Speedy Moisture Tester The speedy moisture tester is a commonly used system that measures the moisture content of unbound materials during construction of embankment and pavement layers. It consists of a rugged plastic case containing a low-pressure vessel fitted with a pressure gauge, an electronic scale, steel balls, reagent, and brushes. Figure 36 shows the main components of the speedy moisture tester, which costs about $2,000. The opera- tional principle of this device is based on measuring the amount of gas produced by a reactant material and the free moisture in the soil to determine the soil’s moisture content. The test is performed according to the ASTM D4944-04 standard or AASHTO T217, and it involves placing about 20 g of soil along with an equal amount of the reactant material (calcium car- bide) in the pressure vessel. After the steel balls are added, the vessel is sealed and inverted to bring the reagent and the soil into contact. The vessel is then shaken for 10 s, followed by a rest period of 20 s. This process is repeated for 1 to 3 min depending on the soil type. Shaking is performed to ensure that all of the moisture reacts with the reagent. Steel balls are used to break up any lumps in the soil sample. Once the reaction is complete, the sample weight and the pressure increase inside the vessel are recorded. Because the pressure in the vessel is proportional to the amount of moisture in the sample, the moisture content can be read directly from the calibrated pressure gauge. The speedy moisture tester takes less than 5 min to obtain results. The test is easy to perform and requires minimal oper- ator training. However, there are some limitations of this test. First, it cannot be used for highly plastic clayey soil because this type of soil may not mix thoroughly with the reagent; thus, some moisture remains in the soil. In addition, some soils may contain chemical compounds that may react with the reagent, yielding erroneous results. Because of the small size of materials that can be tested in the speedy moisture tester, this test may not provide accurate moisture content measurement of coarse-grained granular soil (Berney et al. 2011). Inaccurate measurements may also result from the use of old calcium carbide reagent, incomplete breakdown of FIGURE 36 Speedy moisture tester (Berney et al. 2011).

43 The researchers found that the moisture analyzer under- estimated the moisture content compared with the laboratory oven dry method. In addition, they indicated that the small volume size of the tested sample prohibited the use of this device for unbound materials with aggregates exceeding 1 in. in diameter (Berney et al. 2011). Field Microwave Oven The field microwave oven provides a fairly reliable mois- ture measurement in a very short period of time. According to ASTM D4643-08, a soil sample is repeatedly heated and weighed every minute until the readings become steady, which indicates that the sample is completely dry. ASTM specifies that a 700-W microwave oven should be used for this device, a representative soil sample is placed on a small, disposable aluminum foil dish, and the dish is transferred inside the heating unit. After taking the initial mass, heating of the soil samples is continued until the mass reaches a constant value. The difference between initial and final mass is taken as the mass of water. The gravimetric moisture content is then calculated from mass of water and mass of dry soil sample. The moisture analyzer costs about $1,840 (Sebesta et al. 2013). Synthesis of Previous Studies Limited studies were conducted to evaluate this device. Berney et al. (2011) compared the moisture contents obtained using the moisture analyzer with those obtained using the labora- tory oven dry method. This comparison is shown in Figure 40. FIGURE 37 Moisture content measured by speedy moisture tester and oven dry method (Oman 2004). FIGURE 38 Moisture content measured by speedy and oven dry methods (Berney et al. 2011). FIGURE 39 Sartorious Model MA 150 1,200-g moisture analyzer (Berney et al. 2011).

44 is capable of measuring volumetric moisture content and bulk density of soil samples; therefore, gravimetric moisture con- tent also can be calculated. The DOT600 costs about $3,000 (Sebesta et al. 2013). The device consists of a sample chamber 3 in. in diam- eter that is retrofitted with a waveguide containing interlaced circuit traces that form a capacitor. The waveguide floats on precision springs. The device’s hardware generates and mea- sures a scaled oscillation resonant frequency. Magnetic linear sensors measure sample mass and volume to allow for the determination of gravimetric moisture content. The oscillation testing. Smaller or larger ovens can be used, but care should be taken to avoid over-drying the soil. Drying times do not have a linear relationship with the wattage of the micro- wave oven. Rather, drying times decrease exponentially with increased wattage. Because a microwave continues to add energy to the sample, it can drive outbound water in clay min- erals if not used according to the ASTM standard, resulting in higher measured values for moisture content. On the other hand, internal studies by the U.S. Army Corps of Engineers have shown that the microwave does not dry out the bound mineral water in gypsum and calcium carbonate soils, making it a superior option than the laboratory oven for those types of soils (Berney et al. 2011). The field microwave oven method is suitable for materials consisting of particles smaller than 4.75 mm. Care should be taken when testing materials with larger particles because of the increased chance of particle shattering (ASTM D4643-08). The field microwave is dependent on a constant battery source. If the device is not fully charged, the wattage will decrease as the battery loses charge. This could cause considerable error in the measurement (Berney et al. 2011). Synthesis of Previous Studies In a study conducted by Berney et al. (2011), the field micro- wave oven slightly underestimated moisture contents of the tested soils compared with the laboratory oven (Figure 41). In addition, the accuracy of this device deviated for high mois- ture content silts and clays because of drying of the bound- mineral water in the soil, a water barrier that is not evaporated under constant thermal energy found in a laboratory oven. DOT600 Roadbed Water Content Meter The DOT600 Roadbed Water Content Meter (Figure 42) is a portable moisture measuring device in which soil samples collected from field sites are compacted and water content is measured using dielectric permittivity methods. This device FIGURE 40 Moisture content measured by moisture analyzer and oven dry methods (Berney et al. 2011). FIGURE 41 Moisture content measured by field microwave oven and laboratory oven dry methods (Berney et al. 2011). FIGURE 42 DOT600 Roadbed Water Content Meter (Sebesta et al. 2013).

45 rate measurement) and is ideal for measuring forest soils and road subgrade. When the probe is pressed firmly against the material to be tested, the device emits a small electrical current. The dielectric permittivity and conductivity values are calcu- lated as the current moves through the soil between electrodes on the probe (Camargo et al. 2006). Then the moisture content of the soil is calculated from the measured dielectric permittiv- ity (Camargo et al. 2006). The Percometer costs $7,735. Synthesis of Previous Studies Dai and Kremer (2006) reported that the Percometer was mod- erate in terms of convenience of use. Portability was reported to be moderate as well because the device has no casing. The device’s readings were found to be inconsistent if the soil sur- face was rough and voids on the surface were not completely filled. This tendency was more pronounced with reclaimed materials. The researchers also found that although the dielec- tric constant increased as the moisture content increased, data were scattered, indicating high variability in test results. Trident Moisture Meter The Trident moisture meter uses dielectric permittivity to determine the moisture content of sand, gravel, crushed stone, and other fine and coarse aggregate. The Trident uses its five- pronged sensor, shown in Figure 44, to measure the complex dielectric constant of the material encompassed by the outer four prongs. The manufacturer recommends taking an aver- age of five to 10 readings to ensure an accurate measurement. The integrated microprocessor converts the dielectric constant frequency of the circuit decreases as moisture content of the sample increases. This frequency is empirically calibrated to obtain moisture content. The manufacturer provides calibra- tion of this device using soils with a range of textures. The DOT600 is completely portable, easy to use (minimal training is required), and quickly tests soil moisture content (takes only about 90 s). One of the major limitations of the device is that it cannot be used to test coarser materials. The accuracy of the moisture measurements can be affected by soil type and by soil salinity. Synthesis of Previous Studies Minnesota DOT studied the accuracy and effectiveness of the DOT600 for measuring soil moisture content (Minnesota DOT 2012; Hansen and Nieber 2013). In this study, DOT600 measurements were compared with those taken using standard Proctor test for 270 samples. The optimum moisture contents obtained using the DOT600 were consistent with the measure- ments determined during the standard Proctor test. In addition, researchers found that the variability of moisture measure- ments using DOT600 was far less than the variability in the optimum moisture constant determined during the Proctor test. Based on overall performance, the Minnesota DOT found that the DOT600 can replace the NDG as well as the sand cone and Proctor tests. However, they suggested some modifica- tions before it could be considered a viable alternative, such as making the device rugged enough for regular field use. Percometer The Percometer estimates the moisture content of soils by mea- suring the dielectric permittivity and conductivity. As shown in Figure 43, the device consists of a 6-cm diameter probe attached to a processor. The probe is designed for insertion into soft materials (a minimum depth of 10 cm is required for accu- FIGURE 43 Percometer (Roadscanners 2006). FIGURE 44 Trident handheld microwave moisture meter (James Instruments, Inc. 2012).

46 SUMMARY This chapter reviews the density-based compaction control methods that have been used by state DOTs. It also provides a comprehensive evaluation of non-nuclear devices used to measure density and/or moisture content based on the results of investigations reported in the literature. Despite that they are used by the majority of DOTs in determining the target density value, the AASHTO T99 and AASHTO T180 stan- dards cannot be used for unbound granular materials that have more than 30% by mass of their particles with sizes greater than 19 mm (¾ in.). Furthermore, past research stud- ies indicated that there are other types of laboratory compac- tion tests that might be more suitable for granular unbound materials because they may provide better replicates of field densities. Although previous studies indicated that non- nuclear devices have some advantages over the NDG, such as not requiring special licensing to operate, they were found to be more difficult to operate and require longer testing time. Finally, there are several non-nuclear devices that can mea- sure moisture content; however, only limited studies have been conducted to evaluate most of them. to moisture content value and displays it as a percentage of dry weight. Material-specific calibration is required for highest accuracy. This device is completely portable, and testing is easy and fast. It provides accurate moisture content measurement and has good data storage capability. However, it requires material- specific calibration and cannot be used for aggregates with a size greater than 25 mm. The Trident costs about $1,900. Synthesis of Previous Studies The Trident moisture meter is relatively new, and only a lim- ited number of studies on its use can be found in the literature. Jean-Louis and Gabriel (2010) investigated the correlation between the measurements taken with the Trident and the actual oven moisture contents of selected soil at different levels of compaction. Test results indicated that the Trident moisture content measurements had a strong correlation with those obtained using the traditional oven dry method for both compacted and loose soils.