Silica ume in Shotcrete by John Wolsiefer, Sr., and Dudley R. Morgan

28 Shotcrete • Winter 2003 Table 3 - Plastic properties of wet-mix shotcrete Table 1 - Wet-mix shotcrete mix Mix A B c D designs, kgfm3 Mix type PC USF CLDSF CHDSF Ambient temperature, C 9 10 13 14 Mix A B D c Shotcrete temperature, C 14 12 15 13 Mix type PC USF CLDSF CHDSF Slump, mm Portland cement, Type I 401 350 353 359 Base shotcrete 40 50 45 100 Silica fume - 47 48 46 After SF & superplasticizer - 50 35 20 Coarse aggregate, 10 mm, 462 485 475 467 Air content, percent SSD Base shotcrete 8.5 7.2 8.0 7A Concrete sand, SSD 1258 1213 1239 1263 After SF & superplasticizer - 6A 5.8 5.8 Water 171 177 177 176 As-shot 4.8 3.9 3.2 2.6 Water-reducing 887 1952 1952 1922 admixture, ml Thickness to bond break Superplasticizer, m1 - 1597 1597 1360 Overhead application, mm 95 130 280 180 Air-entraining admixture, 118 296 296 296 Vertical application, mm 305 330 380 405 ml Overhead rebound, percent - 12.9 12.3 10.4 Total 2294 2297 2296 2314 Vertical rebound. nercent 3.4 2.7 3.7 3.9

Table- 2 Dry-mix shotcrete mix Table 4 - Plastic properties of dry-mix designs, kg/m3 shotcrete Mix E , F G H Mix E F .G H Mix type PC USF CLDSF CHDSF Mix type PC USF CLDSF CHDSF Portland cement, Type I 425 373 373 373 Ambient temperature, C 6 6 8 7 Silica fume - 49 49 49 Shotcrete temperature, C 14 16 14 13 Coarse aggregate, 10 mm, 495 491 491 491 Thickness to bond bread SSD Overhead application, mm 65 380 280 230 Concrete sand, SSD 1216 1204 1204 1204 Vertical application, mm 205 460 560 460 Water (estimated) 163 165 165 165 Overhead rebound, percent 42.7 20.4 25.2 18.6 Total 2300 2281 2281 2281 Vertical rebound percent 45.4 21.1 22.9 24.6

Shotcrete test program Mix designs and supply water reacting with the cement and sil­ ica fume is too short for effective A study was undertaken to evaluate the The wet- and dry-mix shotcrete mix water reduction before the mix is actu­ performance characteristics of three designs used are shown in Tables 1 ally consolidated in place on the shot­ different silica fume product forms in and 2. These mix designs are typical of crete surface. both wet-mix and dry-mix shotcrete: those used in rock slope stabilization The wet-mix shotcrete was brought and tunnelling projects in the United • as-produced uncompacted silica to the field test site by transit truck, States and Canada. The cement was a fume (USF) with the silica fume and superplasti­ portland Type I, with aggregates meet­ • compacted low-density silica fume cizer added on-site. A shotcrete piston ing the requirements of the ACI Stand­ (CLDSF) pump was used to apply the wet-mix ard Specification for Materials, • compacted high-density silica fume shotcrete. The dry-mix shotcrete was Proportioning, and Application of (CHDSF) weight-hatched in premixed super Shotcrete, ACI 506.2, Gradation No. 2. The performance characteristics sacks with cement, aggregate, and sil­ The control mixes are labelled A (Wet) ica fume all premixed. The dry-mix evaluated included rebound loss, thick­ and E (Dry). The silica fume mix de­ ness to bond breaking (sloughing) on was premoisturized to a moisture con­ signs, prepared with USF, CLDSF, and tent of 3 to 4 percent prior to discharge overhead and vertical surfaces, com­ CHDSF are designated, respectively, pressive strength, flexural strength, in a rotating barrel feed shotcrete gun. B, C, and D for the wet-mix and F, G, drying shrinkage at 50 percent relative and H for the dry-mix shotcretes. humidity, chloride permeability, elec­ The silica fume dosage averaged 13 Thickness to bond break and trical resistivity, boiled absorption, and percent (by mass of cement) for all sil­ rebound loss volume of permeable voids. These pa­ ica fume shotcrete mix designs. A Silica fume addition to shotcrete in­ rameters were compared to the per­ naphthalene sulphonate-based super­ creases adhesion to the bonding sur­ formance of a shotcrete control mix plasticizer was used to control the face and cohesion within the shotcrete; prepared with plain portland cement water-cement ratio in the wet shotcrete consequently, the thickness of shot­ mix. Superplasticizer is not required crete build-up attainable on overhead for dry-mix shotcrete, since most of and vertical surfaces is substantially *Morgan, D. R., "Recent Developments in Shotcrete Technology," Materials Engineering Perspective pre­ the water in the mix is added at the improved. There is no standard ASTM sented at the World of Concrete 1988, Las Vegas. shotcrete nozzle; contact time for the or ACI test to measure attainable

Shotcrete • Winter 2003 29 84 84 77 77 70 70

"'- 63 63 1§. ~ ;: 56 ;: 56 "' 49 "'~ 49 ....~ ....a: 42 42 ~ "'~ "'> :> ;;; 35 ;;; 35 ~ ~ a: "'- 28 :'!: 28 :E :E ...,0 21 o A - Control Mix. Portland Cement ...,0 21 o E - Control Mix. Portland Cement + F - Uncompacted Silica Fume + B - Uncompacted S~ica Fume 14 o C - Compacted Low Density Silica F ...e 14 0 G - Compacted Low Density Siliea Fume • D - Compacted High Density Silica Fume 0 H - Compacted Hi9h Denaity Silieo Fume 7

0+-----.----.-----r----~----.-----.---~ 0+----.-----.-----r----.----.------~ 0 20 40 60 0 20 40 60 AGE !DAYS) AGE (DAYS) Fig. 1 -Compressive strength of wet-mix shotcrete. Fig. 2 - Compressive strength of dry-mix shotcrete.

Table- 5 Hardened properties of Table 6- Hardened properties of wet-mix shotcrete dry-mix shotcrete Mix ASTM A B c D Mix ASTM A B c D Mix type test PC USF CLDSF CHDSF Mix type I···· test PC USF ... · CLDSF CHDSF procedure procedure Compressive strength, c 39 Compressive strength, C39

MPa MPa ·-· 24 hours 14.5 21.7 16.8 17.3 24 hours - - 24.7 23.7 7days - 44.4 38.6 35.1 29 hours· 31.1 33.8 - - 28 days 43.8 63.5 55.9 57.4 7 days 44.2 49.2 45.2 44.4 63 days 44.0 69.7 64.0 64.9 28 days 53.8 59.9 58.7 54.9 Flexural strength, MPa C78 63 days 61.8 67.2 66.3 62.4 7 days - 4.9 3.8 4.1 Flexural strength, MPa C 78 28 days 5.3 6.7 6.0 6.5 28 days 7.4 8.4 6.6 7.5 Boiled absorption, I C642 5.9 6.6 6.9 6.3 Boiled absorption, C642 4.9 2.7 3.6 4.0 percent, 28 days percent, 28 days Vo!lll1le of permeallie 12.9 14.3 14.9 13.9 Volume of permeable 11.2 6.3 8.3 9.2 voids, percent, 28 days voids, percent, 28 days Bulk specific gravity 2.296 2.304 2.307 2.341 Bulk specific gravity 2.380 2.398 2.371 2.370 after immersion and after immersion and boiling boiling

thickness build-up, so thickness to vertical rebound was reduced from Compressive and flexural bond break (sloughing) and rebound 45.5 percent in the plain control mix to strength loss were measured in a specially con­ 22.8 percent, on average, for the three Compressive strength was measured at structed rebound chamber. These pa­ silica fume product forms. The wet­ 24 hours, and 7, 28, and 63 days by rameters are shown in Tables 3 and 4. mix shotcrete rebound percentages testing cores extracted from shotcrete In the wet-mix shotcrete study, the were low in all mixtures. test panels. The panels were cured in overhead thickness at bond-break was the field for the first 24 hours, then 3.5 in. (90 mm) for the plain portland In summary, the wet-mix data vari­ transferred (in the wooden forms) to a cement Mix A, and reached a maxi­ ance for the three silica fume product laboratory, where the shotcrete was mum of 11 in. (280 mm) in Mix C forms shows no significant difference moist-cured. The strength data shown (CLDSF). The overhead thickness at in rebound loss, but some differences in Table 5 and Fig. I show that using bond break was typically greater for in thickness to bond break. For the silica fume generated significant in­ the dry-mix shotcrete, reaching a dry-mix shotcrete, there is a greater creases in the wet-mix shotcrete com­ pressive strength. The control mix maximum of 15 in. (380 mm) in Mix F thickness of 15 in. (380 mm) for USF compressive strength was 6390 psi (44 compared to 11 and 9 in. (280 and (USF), compared to 2.5 in. (65 mm) MPa) at 63 days compared to an aver­ for the plain Mix E. The dry-mix shot­ 230 mm) for the CLDSF and CHDSF age of 9590 psi (66.1 MPa) for the sil­ crete overhead rebound was decreased mixes, respectively. However, note ica fume shotcretes, about a 50 percent from 42.7 perclent for the plain control that these thicknesses were attained in increase. to an average of 21.4 percent for the a controlled test environment, and may The dry-mix silica fume shotcrete three silica fume product forms. The not be achievable in field applications. compressive strengths also were

30 Shotcrete • Winter 2003 7000 (/) ~ 3000 6000 3 2578 u 2500 5000 >- t: 2000 E - Control Mi•. Portland Cement 4000 A - Confrol Mi>c. Portland Cement F - Uncompacted s•co Fume B - Uncompacted Sica Fume ~ G - Compacted Low Density Silica Fume C - Compacted Low Density Silica Fume H - Compacted High Denoity Silica Fume D - Compacted Hi~ Oenaity Silica Fume 1500 3000 i 1000 2000 ~ ~ 1000 Q 500 ~ a:

MIX DESIGN TYPE MIX DESIGN TYPE Fig. 3- Rapid chloride permeability of wet-mix shotcrete. Fig. 4- Rapid chloride permeability of dry-mix shotcrete.

higher than that of the control mix, though not as pronounced as in the Table 7- Chloride permeability based on charge passed wet-mix shotcretes (Table 6 and Fig. i Charge passed, coulombs Chloride Typical of 2). I permeability The flexural strength specimens Greater than 4000 High High water-cement ratio (0.6) conventional PCC* were cut from the shotcreted panels for 2000 to 4000 Moderate Moderate water-cement ratio (0.4 to 0.5) conventional PCC* 28-day testing. The silica fume wet­ lOOOto 2000 Low Low water-cement ratio (0.4) conventional PCC* mix shotcretes were also tested at 7 100 to 1000 Very low Latex-modified concrete, silica fqme concrete (5 to 15 I percent) .· c. days. The flexural strength data is shown in Tables 5 and 6 for the wet­ Less than I 00 i Negligible IPolymer-impregnated concrete, polymer concrete, and high silica fume content concrete (15 to 20 percent) mix and dry-mix shotcretes, respec­ *Portland cement concrete. tively. The greatest strength improve­ ment is again in the wet-mix silica Rapid chloride permeability and values are in the "high" and "moder­ fume shotcretes. electrical resistivity ate" classification, respectively, for In summary, with respect to com­ concrete,3 as shown in Table 7. Based Chloride permeability and electrical pressive and flexural strength of the on historical data, concrete of this resistivity data were generated from quality would have inferior durability hardened shotcretes, there generally cores cut from the shotcrete panels. performance in an aggressive chloride are only small differences in perform­ Tests were conducted to the require­ environment. In contrast to this data, ance between shotcretes made with the ments of the "Standard Method of Test the silica fume shotcrete reduced the three different silica fume product for Rapid Determination of Chloride forms. Permeability of Concrete," AASHTO chloride permeability to an average of 371 coulombs for the wet-mix shot­ Designation T277-83. Chloride perme­ Boiled absorption and crete and 192 coulombs for the dry­ ability and electrical resistivity are permeable voids mix shotcrete. very important characteristics in evalu­ The boiled absorption, volume of per­ ating the ability of shotcrete in a reha­ The electrical resistivity measure­ ments (Fig. 5 and 6) show correspond­ meable voids, and bulk specific grav­ bilitation application to slow down or ity were measured after immersion and prevent corrosion of steel reinforce­ ingly large improvements over the boiling according to ASTM C 642 test ment. control shotcrete. The dry-mix silica fume shotcrete shows an average elec­ procedures. The data are presented in The rapid chloride permeability data trical resistivity of 55,290 ohms-em, Tables 5 and 6 for wet- and dry-mix are shown in Fig. 3 for wet-mix shot­ shotcretes, respectively. crete and in Fig. 4 for dry-mix shot­ compared to the control mix value of 5490 ohms-em. In this test, silica fume addition re­ crete. In spite of the fairly good All three forms of silica fume in sulted in significant reductions in strength, absorption, and permeable both wet and dry-mix shotcrete result boiled absorption and permeable voids void data for the plain portland cement in chloride permeability reduction 10 in dry-mix shotcrete, but not in wet­ shotcrete control, the rapid chloride to 20 times greater than that mix shotcrete. All the wet-mix shotcre­ permeability was 6800 coulombs for of the control portland cement shotcrete (Fig. tes have absorption and permeable the wet-mix shotcrete and 2573 cou­ lombs for the dry-mix shotcrete. The 3 and 4). This observation, together voids test results that can be rated as with the electrical resistivity data, is a being between "good" and "excellent," very significant indication of the bene­ with all the dry-mix shotcrete data ex­ fits of using silica fume in shotcrete tremely low, being in the "excellent" *Morgan. D. R., "Recent Developments in Shotcrete Technology," Materials Engineering Perspective pre­ for rehabilitation of reinforced con­ category.* sented at the World of Concrete 1988, Las Vegas. crete structures containing deteriorated

Shotcrete • Winter 2003 31 0.14

O.IJ A - Control Mix. Portland Cement ] 40 000 B - Uncompacted Silica Fume 0.12 (/) C - Compacted Low Density Sitlco Fume D - Compacted High Density Silica Fume 0,11 0.1 ~ 34 280 0.09 30 000 ~ >- 0.08 ~ z: ~ "' ~ 0,07 !1l "'u""' (/)w "'.... 0.06 20 000 "' 0.05 "' ~ 0.0< ~ o A - Control Mix. Portland Cement ~u"' O.OJ + B - Uncompacted Silica Fume o C - Compacted Low Density Sdico Fume w 10 000 0.02 o:l 6 D - Compacted High Density SiUca Fume 0.01

20 40 60 80 100 MiX DESIGN TYPE AGE (DAYS! Fig. 5 -Electrical resistivity of wet-mix shotcrete. Fig. 7- Drying shrinkage of wet-mix shotcrete. 0.14 ...------, O,IJ E - Control Mix. Portland Cement F - Uncompacted SHica Fume 70 000 64 120 0.12 G - Compacted Low Density Si~co Fume 'B H - Compacted 1-tgh Density Saico Fume 0.11 60 000 0.1

~ ~ 0.09 w ,.. 50 000 0.08 t= "' 2: ""~ 0,07 40 000 '"'= w~ .... 0.06 a: "' ..J 0.0.5 < 30 000 ~ ~ 0.0< t; w 20 000 O.OJ o E - Control Mix. Portland Cement ;:rl 0.02 + F - Uncompacted Silica Fume 0 G - Compacted Low Density Silica Fume 10 000 0.01 6 H - Compacted High Density Silica Fume 0 20 40 60 80 100 t1X DESIGN TYPE AGE (DAYS) Fig. 6 - Electrical resistivity of dry-mix shotcrete. Fig. 8- Drying shrinkage of dry-mix shotcrete.

steel in environments with chloride ex­ 2. With respect to the wet-mix shot­ forms of silica fume evaluated were posure. crete process, incorporating silica effective in reducing rebound. This has fume in the mix resulted in significant significant cost implications for the increases in achievable thickness of shotcrete process as significant sav­ Drying shrinkage build-up compared to plain portland ings can be achieved by reducing ma­ Drying shrinkage tests were con­ cement shotcrete. The greatest thick­ terials costs and enhancing productiv­ ducted in accordance with ASTM C ness of build-up on overhead surfaces ity. 341 test procedures using specimens was achieved with the compacted low 5. Substituting silica fume for port­ cut from the shotcreted panels. At 56 density silica fume mixture (CLDSF). land cement resulted in modest in­ days, the data show that the uncom­ Rebound was low in all the wet-mix creases in compressive and flexural pacted silica fume shotcrete Mixes B shotcretes studied, with little differ­ strength in dry-mix shotcrete and sub­ and F had the lowest values of drying ence in rebound between the various stantial increases in compressive and shrinkage (Fig. 7 and 8). The dry-mix mixtures evaluated. flexural strength in wet-mix shotcrete. shotcrete shrinkage was lower than for 3. With respect to the dry-mix shot­ Differences in strength attributable to the wet-mix shotcrete, and can be best crete process, incorporating silica explained by the dry-mix shotcrete's fume in the mix resulted in substantial the various forms of silica fume stud­ lower water demand. increases in achievable thickness of ied generally were small. build-up compared to the plain port­ 6. Significant reductions in the val­ land cement shotcrete. The greatest ues of boiled absorption and volume of Summary and conclusions thickness of build-up on overhead sur­ permeable voids were evident in the 1. This study has demonstrated that all faces was achieved with the uncom­ silica fume mixes compared to plain three forms of silica fume studied (un­ pacted silica fume (USF). portland cement for the dry-mix shot­ compacted, compacted low density, 4. Approximately 50 percent reduc­ cretes, but not for the wet-mix shotcre­ compacted high density) can be read­ tion in rebound in dry-mix shotcretes tes. However, the measurement data ily batched, mixed, and applied in both applied to vertical and overhead sur­ for these parameters would place all the dry- and wet-mix shotcrete proc­ faces was achieved by incorporating shotcrete in a "good" to "excellent" esses. silica fume in the mixture; all three category.

32 Shotcrete • Winter 2003 for field placements of up to 18,000 psi concrete. By 1981, high strength and decreased permeability were becoming Reproduced with permission from the April familiar to knowledgeable researchers, 1993 edition of Concrete International— the magazine of the American Concrete Institute. but the first silica fume concrete studies to prevent rebar corrosion were initiated and developed by him and corrosion John Wolsiefer, Sr., is President of specialist Kenneth C. Clear. From this work, Norchem Concrete Products, a silica the evolution of silica fume to enhance fume manufacturer. He has 23 years’ concrete durability through the use of experience, including past CEO positions high-performance concrete has become in consulting, ready mix and concrete the largest single material application. construction. He has a BS in applied physics and an MS in management Dudley R. (Rusty) engineering. He is a member of ACI and Morgan is Chief serves on ACI Committees 234, Silica Materials Engineer Fume in Concrete; and 363, High with AMEC Earth & Strength Concrete; and is the Chair of Environmental Ltd. the ASTM Silica Fume Specification He is a civil engineer Committee. He has given numerous with over 35 years’ presentations, authored papers on silica experience in concrete and shotcrete fume, and received the Asbjorn Markstad technology and the evaluation and Award for significant contributions to rehabilitation of infrastructure. Morgan the development of silica fume concrete is a Fellow of the Canadian Academy of technology. In 1975, as technical consultant Engineering and the American Concrete to the Norwegian Cement Industry, Institute (ACI), and he is Secretary of Wolsiefer initiated the first United States ACI Committee 506, Shotcreting. He is evaluation program to determine the a member of several ACI, ASTM, and positive properties of silica fume concrete Canadian Standards Association (CSA) and the appropriate market applications. technical committees, and he is a founding In 1976, he conducted the first commercial member of the American Shotcrete placements of silica fume concrete for Association. Morgan has provided bulk chemical storage warehouses in the consulting services on concrete and United States and Canada. In 1978, he shotcrete projects throughout North developed silica fume concrete mixes America and around the world.

Shotcrete • Winter 2003 33