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.
83 Appendix B. Details of Laboratory Measurements State Practices State transportation agencies (e.g., DOTs) currently use a wide variety of electrochemical test standards for characterizing the steel corrosion potential of earthen materials. We performed a review of state DOT standard specifications and other state publications for information regarding corrosion of MSE reinforcements. A summary of state DOT practices is presented in Table B-1. Twenty-two of the states were found to generally follow AASHTO requirements and use AASHTO test methods, three states referenced multiple test methods, one state publishes modifications to the AASHTO methods, 12 states do not use the AASHTO methods, 15 states use different electrochemical requirements, and one state uses FHWA guidance instead of AASHTO as a reference in the state specifications. A state may be included in multiple categories, for instance California uses different test methods and different electrochemical requirements and is included in both counts. Table B-1. Summary of state practices. Reference tests States States using AASHTO tests and requirements Alaska, Indiana, Iowa, Maine, Michigan***, Mississippi, Montana***, Nebraska, Nevada**, New Jersey, New York (MSE Inspection Manual), North Carolina (Fine Aggregate Only), North Dakota***, Ohio**, Oklahoma*, Rhode Island***, South Dakota***, Vermont***, Virginia, Washington D.C.***, West Virginia**, Wisconsin* States referencing multiple methods Connecticut****, Delaware, South Carolina States using modified AASHTO test methods Illinois States using different test methods Arizona, California, Colorado, Delaware, Florida, Kansas, Kentucky, Louisiana, New York (Stand. Specs.), Pennsylvania***, Texas, Washington** States using different electrochemical requirements Arizona, California, Colorado, Florida, Georgia*, Illinois, Kansas, Kentucky, Louisiana, Missouri, New Mexico, New York (Stand. Specs.), South Carolina, Texas, Washington States using FHWA-NHI-09-087 Minnesota * test methods not stated explicitly ** Chloride and sulfate tests are waived if resistivity is greater than 5000 â¦-cm *** AASHTO stated as design standards **** Ranges not stated Summary of Resistivity Tests Table B-2 shows the summary of resistivity tests used in this research in terms of specimen preparation, precision (µcov and Ïcov) and bias (µ and Ï) with respect to the results of AASHTO T-288 (2016).
84 Table B-2. Summary of resistivity tests, precision, and bias. Test Salient details Precision Bias w.r.t. AASHTO T-288 Soil box tests Measure resistivity (Ï) of compacted specimens µcov Ïcov µ Ï AASHTO T-288 (2016) Sample passing No. 10 sieve, incremental addition of water (10% by weight), curing for 12 hours, compacting the specimen into the soil box and measuring the resistance. Keep adding water until reaching Ïmin or Ïsat 4.6 4.51 - - ASTM G- 187 (2018) Scalp material greater than ¼ inch, saturating the specimen while compacting in a soil box and measuring resistivity at saturation, Ïsat. Saturation is determined visually 5.3 4.5 1.4 0.4 ASTM WK 24621 Include particle sizes up to 1 ¾ inches, curing for 24 hours, compacted specimen into the soil box. The specimen is allowed to drain before making the resistance readings 7.4 5.8 3.8 5.1 Tex-129-E (1999) Similar to AASHTO T-288 except the material is separated on a No. 8 sieve before testing 3.2 4.3 1.1 0.2 Tex-129-M Include particle sizes up to 1 ¾ inches, sample is removed from the box after each measurement and re-mixed with additional water, test continues until the additional moisture cannot be absorbed and the specimen has reached either Ïmin or Ïsat 4.8 2.8 2.3 1.9 Leaching tests Measure conductivity (κ) of aqueous solution µcov Ïcov µ Ï SCT 143 (2008) Include particle sizes up to 1 ¾ inches, diluting the sample at a 1:1 ratio, agitating for 3-minutes at the 30- minute, 2-hour and 4-hour intervals, standing for 20 hours and filtering the sample before conductivity measurements 4.9 4.5 1.9 1.7 Tex-620-J (2005) Sample passing No. 40 sieve, diluting the sample at a 10:1 ratio, heating to 140°F while stirring every hour for 12 hours, measuring conductivity of the mixture 4.9 1.9 1.6 0.9 Tex-620-M Include particle sizes up to 1 ¾ inches, diluting the sample at a 10:1 ratio, mixed for 60 minutes, not allowed to stand before the conductivity measurements 4.5 3.1 5.2 3.4 1 In this study precision for AASHTO 288 was determined using 100% saturation as the end point.
85 Comparison of Different Resistivity Tests with AASHTO T-288 Table B-3. Statistics of resistivity test bias with respect to AASHTO T-288 (2016). Test method µbias Ïbias COVbias (%) = Ïbias/µbias Tex-129-E 1.07 0.24 22 ASTM G-187 1.41 0.44 31 Tex-129-M 2.28 1.96 86 ASTM WK 24621 3.75 5.10 136 Tex-620-J 1.69 0.96 57 SC T-143 1.95 1.72 88 Tex-620-M 5.22 3.41 65 Table B-4. Bias of resistivity measurements from samples with different texture. Test method Fine sand Coarse sand Gravel µbias COVbias (%) µbias COVbias (%) µbias COVbias (%) Tex-129-E 1.0 11 1.2 18 1.0 26 ASTM G-187 1.0 12 1.6 25 1.4 35 Tex-129-M 0.9 3 1.6 56 3.1 77 ASTM WK 24621 0.9 6 2.0 40 5.8 116 Tex-620-J 1.3 56 1.8 58 1.7 59 SC T-1431 - - 0.9 96 2.6 72 Tex-620-M 2.2 44 4.6 72 6.5 53 1could not perform SCT 143 (2008) with fine samples due to the lack of settlement of the finer particles during the specified standing time.
86 Summary of Tests to Measure Salt Contents Table B-5 shows the summary of tests used in this research to determine sulfate and chloride contents. This table summarizes each test in terms of specimen preparation and precision COV (µcov and Ïcov). Table B-5. Tests for measurements of salt content, and observations of precision. Sulfate content Test Salient details Precision1 COV (%) µcov Ïcov AASHTO T- 290 (2016) Sample passing a No. 10 sieve, diluting the sample at 3:1 ratio, mixing in a flask, centrifuging and filtering the sample and measuring sulfate ion content using spectrophotometry or ion chromatography (IC) 10.1 5.6 Tex-620-J (2005) Sample passing a No. 40 sieve, diluting the sample at 10:1 ratio, heating to 140°F while stirring every hour for 12 hours and measure sulfate and chloride ion contents via IC 11.8 7.6 Tex-620-M Include particle sizes up to 1 ¾â, diluting the sample at 10:1 ratio, mixing for 60 minutes, not allowed to stand, filtering before the sulfate and chloride measurements via IC 10.7 5.9 Chloride AASHTO T- 291 (2013) Sample passing a No. 10 sieve, diluting the sample at 3:1 ratio, mixing in a flask for 20 seconds, after one-hour repeat shaking, centrifuging and filtering the sample then measuring chloride content by titration or IC 7.5 5.4 Tex-620-J Sample passing a No. 40 sieve, diluting the sample at 10:1 ratio, heating to 140°F while stirring every hour for 12 hours, filtering the sample and measuring the sulfate and chloride contents via IC 3.7 2.5 Tex-620-M Include particle sizes up to 1 ¾â, diluting the sample at 10:1 ratio, mixing for 60 minutes, not allowed to stand, filtering before the sulfate and chloride contents measurements via IC 12.92 6 1 Only included results from testing sulfate and chloride with > 10 mg/kg 2 Outlier from SC LWF removed
87 Summary of Tests to pH Table B-6 shows the summary of tests used in this research to determine pH of the aqueous solutions. This table summarizes each test in terms of specimen preparation, precision COV (µcov and Ïcov) and bias with respect to AASHTO T-289 (2018). Table B-6. Summary of tests for pH and observations of precision and bias. TEST Salient Details Precision COV (%) Bias w.r.t. T-289 (2018) µcov Ïcov µ Ï AASHTO T-289 (2018) Sample passing a No. 10 sieve, air drying, a 30-gram specimen is taken, diluting the sample at 1:1 ratio, stand for a minimum of 1 hr. while stirring every 10 to 15 minutes then measuring the pH of the slurry specimen 1.0 1.0 - - ASTM D 4972 (2019) Sample passing a No. 10 sieve, air drying, a 10-gram specimen is taken, diluting the sample at 1:1 ratio, mixed thoroughly and standing for 1 hr. then measuring the pH of the partially settled suspension 1.1 0.6 0.99 0.03 NCHRP 21-06 (2009) Particles larger than 3/8 in. are removed from the sample by hand. The moisture content of the sample is determined but the sample is not air-dried, diluting the sample at 1:1 ratio, stirring thoroughly to disperse the soil, standing for 30 minutes then the pH of the supernatant is measured 0.7 0.5 0.98 0.03 Tex-128-E (1999) Sample passing a No. 40 sieve, air drying, adding 200 ml of DI water to the sample, heating to 112°F - 140°F, diluting 30-gram of heated dry soil at 5:1 dilution ratio, stirring initially and every 15 minutes for 1 hour thereafter, not allowed to stand then measuring the pH of the suspension 0.9 0.5 1.03 0.03 Tex-620-J (2005) Sample is air-dried and the portion of the sample passing a No. 4 sieve (crushed as necessary) is pulverized to pass a No. 40 sieve, diluting 30-gram of sample at 10:1 dilution ratio, heating to 150°F and stirring, allowing to digest for 15 to 18 hours, the mixture is not allowed to stand then pH of the suspension is measured 0.9 0.3 0.97 0.04 Tex-620-M Particle sizes up to 1 ¾ inches are included in the specimen, air-drying and diluting at 10:1 ratio, agitating for 60 minutes the mixture is not allowed to stand then pH of the suspension is measured 1.2 1.7 1.1 0.04
88 Salt Contents Table B-7 is a summary of the samples with general descriptions of results from testing via the AASHTO and the Texas modified procedures. Results are described as 1) low salt contents with sulfate and chloride contents < 50 mg/kg; (2) both sulfate and chloride contents > â 50 mg/kg, (3) high sulfate (> 50 mg/kg) and low chloride contents (< 50 mg/kg); and (4) high chloride (> 50 mg/kg) and low sulfate contents (< 50 mg/kg). Table B-7. Summary of salt contents measured via AASHTO and Texas modified procedures. Site Results AASHTO Texas Modified Ocala, Florida Low salt1 Low salt M-U-D, NYS Low salt Low salt Pharr, TX Low salt Low salt Prince George, BC Low salt Low salt Ashdown. AR Low salt Low salt Raleigh, NC Low salt Low salt Temple, TX SO4 (41) + Cl- (22) Low salt MSE Coarse, El Paso, TX SO4 (145) + Cl- (85) Low salt PIP 5â, NYS SO4 (56) + Cl- (172) SO4 (26) + Cl- (65) PIP 10â, NYS SO4 (60) + Cl- (137) SO4 (76) + Cl- (105) PIP 15â, NYS SO4 (71) + Cl- (190) SO4 (97) + Cl- (204) MSE Fine, El Paso, TX SO4 (67) + Cl- (45) SO4 (91) + Cl- (43) Quarry, El Paso, TX SO4 (312) + Cl- (204) SO4 (214) + Cl- (121) Maple Rd., NYS High SO4 (153); Low Cl- (43) Low salt LWF-Crushed, LA High SO4 (155); Low Cl- (14) High SO4 (81); Low Cl- (4) Granular Base, SC High SO4 (280); Low Cl- (6) Low salt Round Rock, TX High SO4 (305); Low Cl- (16) Low salt LWF, SC High SO4 (397); Low Cl- (15) High SO4 (427); Low Cl- (21) Calgary, BC High SO4 (809); Low Cl- (11) High SO4 (652); Low Cl- (9) Sprain Brook, NY Low SO4 (63); High Cl- (341) Low SO4 (51); High Cl- (153) Rochester, NY Low SO4 (45); High Cl- (361) Low SO4 (22); High Cl- (145) Wake, NC NA Low salt LWF Uncrushed, LA NA Low salt Waco, TX NA Low salt Bastrop NA Low salt Garden City, TX NA SO4 (197) + Cl- (113) 1 Low salt = total salt content (sulfate + chloride) is less than 50 mg/kg. Data in Table B-7 indicate the results between the AASHTO and Texas modified test procedures are similar for 16 out of 21 samples. Differences include five samples where data from Tex-620- M indicate low salt contents but data from the AASHTO tests indicate total salt contents greater than 50 mg/kg or high sulfate contents. In each case where differences are observed, the finer sand component of the sample is limited and minimal fine sand is included within the Tex-620-M specimen. Four of the samples with different results are gravels with less than 30 % passing a #10 sieve (Temple, TX; Maple Rd., NY; Round Rock, TX; and El Paso, TX âMSE Coarse) and one is a coarse sand with less than 20 % passing the #40 sieve (South Carolina, GB).
89 Results from testing coarse samples with less than 8 % passing a # 10 sieve via Tex-620-M (but not via AASHTO T-290 and 291) show low salt contents. The sample from Garden City Texas with higher chloride and sulfate ion contents has 22 % passing a #10 sieve. The Garden City sample does not have a corresponding AASHTO test result. Bias Statistics We computed bias as the ratio of âequivalent total salt contentâ obtained from Tex-620-M divided by the âequivalent total salt contentâ computed from the results of AAAHTO T-290 (2016) and T- 291 (2013). Equivalent total salt contents consider the combining power of chloride and sulfate in solution in terms of their milliequivalent units, and is useful to check trends between salt content and resistivity. The milliequivalent of an electrolyte is the mass proportion (mg/kg) divided by the equivalent molecular weight, whereby 1 mEq/kg of Cl- is equal to 0.74 mEq /kg of sulfate (SO4 -2). We convert the chloride and sulfate contents to an equivalent total content in terms of chloride as 0.65 (SO4 (mg/kg)) + Cl- (mg/kg). In general, lower salt contents are measured via the modified test procedures so the bias is less than one. Figure B-1 is a histogram depicting the distribution of the bias. The bias ranges between 0.1 and 2.3, with mean bias, µbias, equal to 0.72 and coefficient of variation (COVbias) equal to 65 percent. The highest bias corresponds to the sample from Ocala, Florida that has a very low salt content (< 10 mg/kg). Low salt contents (<10 mg/kg) are affected by the sensitivity of the measurements, and the measurements are more uncertain compared to when higher salt contents are measured. The lowest values of bias are from samples that have a low percentage of particles passing a No. 10 sieve. Figure B-1 Histogram of bias in salt measurements from Tex-620-M. We identified the trend in bias with respect to percent passing the No. 10 sieve for each sample as shown in Figure B-2. After two outliers are removed including the sample from Ocala, Florida and the sample of granular base from South Carolina, a good correlation (R2 = 0.76) is observed between the remaining 19 data points. The Ocala data point was an outlier due to the low salt content and the South Carolina GB sample exhibits a bias that is too low considering other samples with similar medium to fine sand (% passing No. 10) content. The lowest biases are from samples 0 2 4 6 8 10 0.11 0.55 0.99 1.42 1.86 More Fr eq ue nc y 620-Mcl + (0.65) * 620-MSO4/(0.65* T-290+T-291)
90 with less than 25 percent passing a No. 10 sieve. For samples with more than 60 percent of the particles passing a No. 10 sieve, the bias is close to, or greater, than one. For these samples, the fines and fine sand components dominate the leaching of salts from samples tested via either the AASHTO or Texas modified procedures. Higher dilution ratios and different methods of mixing may render measurements of salt contents from Tex-620-M that are higher compared to AASHTO T-290 (2016) and T-291 (2013) for samples with a large fraction passing a No. 10 sieve. Figure B-2 Correlation between bias in salt measurements fromTex-620-M and percent passing the No. 10 sieve used to characterize the sample. Concepts of Resistivity Bias Arciniega et al., 2018; 2019 at The University of Texas, El Paso (UTEP) developed relationships describing how the resistivity of a soil sample (Ï) can be computed based upon physicochemical characteristics including gradation, degree of saturation, and the leaching potential of cations and anions present on the surfaces of the soil particles. That relationship is of the form: ðð = � ð¤ð¤ðð ðððð + ð¤ð¤ðððð ðððððð + ð¤ð¤ðððð ðððððð + ð¤ð¤ðð ðððð � â1 (B-1) where wg, wcs, wfs, wf, are the percent by weight of the constituents including gravel (g), coarse sand (cs), fine sand (fs) and fines (f), and Ï are resistivities of the constituent geomaterials including gravel (Ïg), coarse sand (Ïcs), fine sand (Ïfs), and fines (Ïf) > 60% Ocala, FL < 25% SCDOT - GB
91 For a degree of saturation equal to 100 percent, the resistivity of a saturated geomaterial (Ïsgm) can be approximated as: ððððððð ð ~ ðððððð ð ð ððððððððð ð (B-2) where Ïpw is the tortuosity of the pore water in the system, κpw is the electrical conductivity of the porewater, and ngm is the porosity of the geomaterial/constituent. The electrochemistry of the porewater solution (κpw) is computed based upon the leaching potential including the molar concentration (ci) of soluble ions including sulfate and chloride ions (i), the electrical charge of the ions (zi eq/mol) and the ionic diffusivity (Di) as follows: ð ð ðððð = ð¹ð¹2 ð¶ð¶ðððð �ð·ð·ððð§ð§ðððððð ðð (B-3) where F is Faradayâs constant (96,485 C/eq), Rg is the molar gas constant, T is the solution temperature, and â is to sum the effects from all ions (i), including cations (e.g., sodium, calcium, magnesium, potassium, etc.) and anions (e.g., sulfate, chloride, bicarbonate, etc.). Because the density of water in these conditions is approximately 1 kg/L, the concentration of each ion is approximately equal to the mass concentration or âsalt contentâ (wi, mg of ion per kg of soil) divided by the dilution factor (DF, kg of water per kg of soil). Most of the soluble ions are extracted (leached) from the surfaces of the geomaterial particles (with negligible contribution by diffusion from within the geomaterial particles, except for very porous materials such as expanded clays), so the amounts of ions that can be extracted generally depend upon the external surface areas of the particles and the surface concentration of ions. If we idealize the solid particles to be shaped as spheres such that the surface area of each particle size (j) is Ïdj2, and if we assume an even surface concentration (Îi, mol/m2) of each ion across all particle sizes, then we can write: ð ð ðððð â 1 ððððððð ð ððððð ð â ð´ð´ ðððð â ð´ð´ ð¤ð¤ðð ð·ð·ð¹ð¹ â ð´ð´ ð¤ð¤ðð ðððð2 (B-4) We can use a characteristic particle size (e.g., dmax, d85, d50, d10) as a scalar representative of the specific-surface-area weighted-average particle size, where dmax is the maximum particle size, and d85, d50, and d10 are particle diameters with 85, 50, and 10 percent passing, respectively. (Note that, as a particle size distribution approaches ideal uniformity, all of these characteristic sizes converge.)
92 Tortuosity, Ïpw, which affects the current path through the soil, depends upon the material gradation, degree of saturation, and the particle shapes, angularity, size and packing/porosity. We assume that tortuosity is inherent to the material characterization as fines, fine sand, coarse sand or gravel. Arciniega et al. (2019) confirmed this assumption and demonstrated fitting parameters for the tortuosity function that varied with respect to characteristics of the gradation. For example, proportionalities may be written for soils tested via Tex-129-M (entire gradation) and for AASHTO T-288 (2016) (fine to medium sand component). Compacted soil specimens tested according to Tex-129-M may include gravel, coarse sand, fine sand and fines. Thus, using Equations (B-1) and (B-2) with the proportionality from Equation (B-4), the resistivity as measured by Tex-129-M is proportional to the percentages of each constituent and their particle sizes as: ðð(129âðð) â 1 ðð(129âðð) �ð¤ð¤ðððððð2 + ð¤ð¤ðððððððððð2 + ð¤ð¤ðððððððððð2 + ð¤ð¤ðððððð2� â1 (B-5) AASHTO T-288 (2016) specifies preparing compacted specimens from the portion of the sample passing a No. 10 sieve. Thus, the resistivity as measured by AASHTO T-288 (2016) is proportional to the effective diameter (d#10, not to be confused with d10) and the porosity of the compacted specimen with 100 percent passing a No. 10 sieve (nT-288) as follows: ðð(ððâ288) â �ðð(ððâ288) à 100 % à ðð#102 � â1 (B-6) The ratio (bias) of the results from Tex-129-M relative to results from AASHTO T-288 is determined from these proportionalities. This serves as a scaling parameter relating the bias of the resistance measurements to characteristics of the samples in terms of maximum particle size, gradation and compaction (porosity). ðð129âðð ððððâ288 â ðð(ððâ288) (100 %) ðð#102 ðð(129âðð)(ð¤ð¤ðððððð2 + ð¤ð¤ðððððððððð2 + ð¤ð¤ðððððððððð2 + ð¤ð¤ðððððð2) (B-7) We are considering the effective diameters to be the maximum diameter of the particles included in the constituent as follows. dg = dmax of the sample dcs â 0.25 inches corresponding to the ¼ inch sieve dfs â 0.017 inches corresponding to the #40 sieve df â 0.0028 inches corresponding to the #200 sieve d#10 â 0.079 inches corresponding to the #10 sieve, and nT-288 is the porosity of the compacted specimen (passing No.10) tested via AASHTO T-288 n129-M is the porosity of the compacted specimen tested via Tex-129-M
93 Figure B-3 depicts the trend of the bias considering results from Tex-129-M compared to those from AASHTO T-288. The scaling factor is determined from each sample and plotted along the x-axis. The data are grouped to distinguish results from tests performed with fine sand, coarse sand or gravel samples. The observed correlations demonstrate that the bias from resistivity measurements have trends that are dependent upon the size of the materials; the trends from testing gravel samples are different from those observed from testing coarse or fine sand, as shown in Figure B-3. The bias of results from Tex-129-M compared to those from AASHTO T-288 is observed to be greater than one, and usually less than three. For fine sand, the bias is close to 1, as expected, and appears to be a tail to the trend observed from coarse sand. For the coarse sand, there is an increase in bias with respect to decreasing values of the scaling parameter (inversely proportional). We observe that for coarse sands, the scaling parameter tends to decrease as the gravel content increases and approaches 50 percent. For the case of coarse sand, gravel sized particles are not in direct contact with one another and are enclosed within a matrix of coarse and fine sand sized particles. Under these conditions, an increase in gravel-sized particles increases the tortuosity of the current path during measurement of resistivity, rendering higher measurements of resistivity for the Tex-129-M test compared to results from AASHTO T-288, and an increase in the bias. For coarse sand, the observed bias has a range mostly between 1 and 3. For the gravel category, bias tends to be directly, rather than inversely proportional to the scaling parameter. For gravel materials, the scaling parameter increases due to the relative increase in the coarse sand (wcs) component. For gravel type material, the gravel particles are in contact and the finer fractions (CS, FS and F) pack the spaces between the gravel particles. The gravel component is packed more tightly when the percent gravel is higher, and the percent of coarse sand and fine sand are relatively lower. As the gravel content of the samples decreases and the coarse sand content increases, the more efficient packing arrangement decreases porosity and increases the tortuosity of the current path. These factors affect a higher measurement of resistivity via the Tex- 129-M test, which tends to increase the bias. For gravel materials, the observed bias ranges between 1 and 7, although most of the observed bias is between 1 and 3.
94 Figure B-3 Correlation of bias from Tex-129-M with respect to scaling parameter. 0 1 2 3 4 5 6 7 8 9 0.001 0.010 0.100 1.000 B ia s Te x- 12 9- M /A A S H TO T -2 88 d2#10*nT288*100/(n129-M*(wg*d2g+wcs*d2cs+wfs*d2fs+wf*d2f)) Fine Sand Coarse Sand Gravel