PLACE IN RETURN BOX to move this checkout from you: record. TO AVOID FINES Mum on or before dds due. ‘ DATE DUE DATE DUE DATE DUE », MSU Is An Affirmdivo ActioNEqml Opponunfly Institution mm pita-9.1 A GEOLOGIC STUDYmust unequivo- cally assure that the advantages to be gained are commensurate with the added expense (Rhoades, p. 1.38). It is of interest to note that commercial application of benefi- cation using a heavy media separation.unit to upgrade gravel materials has been inaugerated in Michigan. The process removes objectionable chert and shale in order to meet Michigan State Highway Department specifications. At the present time three gravel beneficiation plants are in operation in the state. PURPOSE Late in 1950 the Michigan State Highway Department completed a large highway bridge in Lansing which carries a four lane divided high- way across the Grand River, a railroad spur line and a four lane interstate highway. Crushed limestone shipped in from Huron County was used exclusively as the coarse aggregate in the construction of the superstructure of this bridge. Approximately one year after the project was completed excessive scaling, erasing and cracking was noted on the curbs, divider strip, sidewalks and deck. In addition to the above failures a large number of pits and pop-outs were developing. In view of the above premature deterioration and excessive failure the Michigan State Highway Department Research Laboratory, Testing and Research Division, East Lansing, Michigan, conducted a thorough inves- tigation of the bridge itself and the materials used in its construc- tion. Several cores were cut and removed from the divider strip, curbs, deck and sidewalks for testing and study. Believing the failure to be predominantly due to the coarse aggregate used, an investigation of the Bayport limestone was conducted at the same time. This part of the study was carried on not only to help fix the responsibility for the failures of the concrete but also to judge the advisability of further use of Bayport limestone for future Highway Department construction projects. Lacking trained petrographers on their staff the Research Labora- tory, Testing and Research Division, made the samples and test data available for geologic study. The solution to the problem, at first hand, seemed to be simply a thorough petrographic examination of the samples. It soon became apparent that such an examination could not begin to explain the wide diversity of results obtained by the standard acceptance tests except in very general terms. It was Obvious that the several constituents of the multifarious samples must be reduced to measureable terms in order to relate the quanti- tative effect of each constituent to the physical properties of each sample. Lacking precedent in this kind of quantitative correlation, and not knowing precisely which component was responsible for pro— ducing the greatest effect on the inherent physical properties of the samples, it was necessary to isolate each and every integral fraction of the sample in order to relate their individual effects on the total physical characteristics of each specimen. The chief objection to the standard acceptance tests and strength tests made by all engineering laboratories is the expense and especially the time involved to obtain satisfactory and conclusive results. The sulphate soundness tests take at least five days; the freezing and thaw- ing in water test requires twenty-five days; various compression and strength tests involve seven to thirty days. Each of these tests must be carried on in a laboratory with special equipment and trained per- sonnel. No satisfactory field tests have as yet been devised to measure the essential properties of concrete aggregates. It was the hope of the author, and the idea pervaded throughout this investigation, to not only unmistakably relate the effects of the fundamental constituents of limestone to its soundness and absorption, but to do so simply and directly without employing complicated equipment or lengthy proceedure. Reference or evidence of this goal will crop up many times throughout this report. In other words, it is hoped that the results of this study will point the way towards a satisfactory and acceptable field evaluation of concrete aggregates which can be used to evaluate deve- loped deposits and also in the exploration for new sources. PROCEDURE The first part of this investigation, the collection of the samples and the physical testing of their specific gravity, absorp- tion and response to freezing and thawing and to magnesium sulphate soundness tests, was performed by the Michigan State Highwny Depart- ment Research Laboratory, Testing and Research Division, East Lansing, Michigan. Additional strength tests and.measurements of thermal expansion were not performed on the individual Bayport samples so that correlation with these special factors could not be accomplished. Collection of anrgz and Stockpilc Samples. The quarry and stockpile samples used in this study were collected by Highway Department personnel in the latter part of 1951. The Bay- port limestone samples from Huron County were collected from three loca- tions on the exposed quarry face in order to include each prominent ledge of the Bayport (Mississippian) Formation from this area. Representative sampling from top to bottom of each ledge, to simulate channel sampling as much as possible, was undertaken to insure complete representation of each ledge. .5233 3388i 3339.3 —Carbonat e s—m h soenaunenoa H Average Percent Loss .________Average Percent Loss r_____Percent 2h hour Absorption L .N seamen .nmmwnows .msamceg poem .coamfl>«c common tom one wmaemee .hnop immoneg nonsenem pace $538 animal 33m Sewing 23 .3 e303 Inoo memes eonepaeoo< oneenepm .o.m.m.<.<# cowpqaomn< one #s — hafi>mnc owneoomm \H. onecucoom _ epmnoasm adaeonmm: — nacho m _\ house 5 moans—.8. one melanoma—III _ ofimD mm _ eoaaadm mHmMada dUHUOAOg EBMpepHpcmso —Ircawoxoopm was handed .3. fishnna...“ agidhaJ)‘ \uoaisat, The two quarry samples of the Delaware (Devonian) limestone from.Silica, Lucas County, Ohio, were taken from near the top and bottom of the current quarry face as representative of the formation from that location. The remainder of the samples used in this study are ”composite" samples of Bayport limestone, Delaware limestone and also Burnt Bluff (Silurian) limestone from Mackinac County, Michigan. The samples were taken from graded crushed stone stockpiles either at the quarry site or on-the-job locations. Although this investigation was orginally and predominately con- eerned with the Bayport limestone these latter samples were included primarily for comparison. In addition they also serve to increase the size of the statistical sample. Comparatively, the crushed Burnt Bluff limestone from.Mackinac County is considered to be superior concrete aggregate by the State Highway Department. In order to determine the cause of failure of a relatively poor aggregate it was advisable to include an acceptable aggregate for comparative purposes. gagineerigg Standard Acceptance Testing The soundness or durability of concrete aggregates refers to their ability to resist disintegration as a result of weathering, variations in temperature, changes in moisture content and certain chemical reactions. Aggregates which disintegrate under such condi- tions are detrimental to concrete and should be avoided. Soundness tests are especially applicable to aggregate materials from.new sources which may be of doubtful quality. The results of soundness tests are regarded as indicative rather than conclusive (Bateman, p. 65). The principle of the freezing and thawing test is relatively simple, depending on the breakdown resulting from one or more of the following reasons; (1) development of excessive hydrostatic pressure within materials of low permeability during progressive freezing, (Blanks, pp. hZO-l), (2) differential thermal eXpansion or contrac- tion of dissimilar materials, and (3) the expansive force of freezing water or the formation of void filling ice crystals. The freezing and thawing test has been almost entirely replaced by the more rapid magnesium or sodium sulphate soundness tests. The American Society for Testing Materials has withdrawn the freezing and thawing test from their specification standards. However, many organizations have retained the freezing and thawing test to the ex! tent that any aggregate failing the more rigorous sulphate soundness test may be approved if it successfully passes a suitable freezing and thawing test. Time, more equipment, and difficulty in standard- izing the freezing and thawing test have contributed to its abandon- .O.te The mechanism of the sulphate soundness tests differs in some respects from the processes by which freezing and thawing of water cause breakdown. During the drying period precipitated crystals of NazSOmor MgSOh fill the pores of the aggregate material. Upon re- immersing the aggregate the precipitated crystals hydrate to NaZSOL- lOHZO or M330, ° 7H20. The resultant expansion of the crystals create stresses which rupture the walls of the pores and cause breakdown of unsound materials (Mielenz, p. 313). The result of the freezing and thawing test and sulphate soundness tests do not parallel each other. The sulphate soundness tests are much more harsh than the longer freezing and thawing in water test. Specific gravity and absorption are sometimes used as prelimi- nary indications of unsoundness but extreme care should be taken to avoid final appraisal of an aggregate on the basis of these tests alone. For example, chert usually has a very low degree of absorp— tion but is very unsatisfactory as a constituent of concrete. Also a high gravity rock may be very unsound due to deleterious impurities. _2_§ m Freezing ing Thawgg E; W (A.A.S.H.O. Method T 103-12). The freezing and thawing test and the following two tests are defined completely and in exacting detail in the American Association of State Highway Officials ”Standard Sggcifications‘fgg Highway EEE' 221.5%: 259 Methods 95 Sampling 25g Testin , “Part II, 191.2; and/or the American Society for Testing Materials ”_2§2“§ggkmg£ Standards”, Part 3, 1956. Only a brief resume of the procedures will be attemp- ' ted here. In the freezing and thawing test the crushed limestone was washed, dried.and separated into appropriate sizes by screening. In order to reduce the tedious time-consuming procedure involved in handling too many sieve fractions the Testing and Research Division made certain innovations in the standard sieve sizes specified for the test and the amount of aggregate used in each size fraction. Only that material smaller than 1 inch and larger than the openings in the # A sieve sep- arated into proportions which were retained on the Q inch, B/L inch, and 1 inch screens was used in the test. In general, approximately L00 to 700 grams of material was retained in each size with the smaller sizes being proportionately smaller then the next larger sieve fraction. After weighing, each separate size of each sample was placed in a shallow metal tray specified as 2% inches x 18 inches x 25 inches in depth. The samples were immersed in water for 2h hours. The water in excess of that necessary to cover the aggregate by i inch was poured off. Then the trays were nested in a metal container 3 7/8 inches x 18k inches x 20 inches and placed in a suitable freezing chamber. The temperature was maintained between -lO° and.-20° F. After freezing for six hours the trays were submerged in a thawing tank at 80° F. Upon thawing for 18 hours the excess water was again removed and the cycle repeated until 25 cycles had been completed. When the final cycle had.been completed the samples were oven dried and sieved over each original sieve on which the material was retained at the beginning of the test. The loss in weight of each sieve frac- tion was calculated as the percentage loss of the fraction. A weighted average loss based on the grading of the material was computed. This weighted “average percent loss" is recorded in column A of table 2 and is used in the statistical comparison. Michigan State Highway Department specifications do not designate the average percent less allowed for sound aggregate material on completion of 25 cycles of freezing and thawing in water. However, the 12 percent loss allowed with 5 cycles of immersion in magnesium sulphate may be used as a guide. ‘2 chle Magggsium.Sulphate Soundness (A.S.T.M. Designation: 088-39T; also A.A.S.H.0., method T - IDA-£6). The same quantities and methods of preparation are used in this test as were outlined for the freezing and thawing test above. The Michigan State Highway Department Research Laboratory uses magnesium sulphate exclusively because of its superior temperature stability over sodium sulphate. Magnesium sulphate is also preferred because of its greater solubility and its reliability in producing more consistent results. The saturated solution of magnesium sulphate was prepared by diss- olving sufficient salt in water at 77° to 86° F. The solution was then cooled to 70° F. and maintained at that temperature for L8 hours before use. The aggregate was placed in wire baskets for ease in draining and immersed in the solution of magnesium sulphate for 16 to 18 hours at 70° F. The samples were then drained and dried to a constant weight in an oven with the temperature at from.221° to 230° F. When thorou- ghly dried the aggregate was cooled for 2 hours at room temperature and then reimmersed in the solution for the beginning of the next cycle. This process was repeated until 5 cycles had been completed. Upon completion of the final cycle the samples were washed free of mag- nesium sulphate and the weighted average percent loss determined in the same way as described under the freezing and thawing sound- ness test. Any aggregate having more than 12 percent loss after 5 cycles of the magnesium sulphate soundness test is classified as unsound for all major highway uses by Michigan State Highway Depart- ment specifications (Mich. State Highway Dept., Table 2, p. ALA). The average percent loss resulting from the magnesium.sulphate soundness test are compiled in Column 3, Table 2, pages 28 and 29. Specific Gravity‘ggg Absorption (A.S.T.M. Designation: C 127-39; also A.A.S.H.O. Method T-85-35.) Specific gravity and absorption are always measured simultanr eously in the same test procedure. Approximately 500 grams of the crushed stone was washed, immersed and allowed to absorb water for 24 hours. The samples were then rolled in a cloth to remove all visible films of water from the particles. The weight of the mat- erial in this saturated surface dry condition.was obtained. The aggregate was then placed in.a wire basket suspended in water to det- ermine its weight in water. Finally, the sample was oven dried to constant weight, cooled to room temperature and weighed. Using these weights the following values were computed; Weight of saturated surface- dry sample in air Bulk 39°31f1c Gm’fl't'yg‘weight of saturated 'Weight of saturated surface-dry sample - surface-dry sample in air in water weight of saturated. Weight of ovens surface-dry sample '— dry sample in in a1 air Percent Absorption - -———-gf,;—18ht of mama”. x 100 sample in air The results for specific gravity and absorption can be found in Columns 1 and 2, Table 2, pages 28 and 29. Geologic Examinatiop and Investigation The geologic procedure followed in this investigation is simple and direct, being nothing more than an enlargement on the usual methods employed in various insoluble residue studies combined with ordinary petrographic examination and the practices of sedimentary petrography. Only the usual equipment found in a geologic laboratory was used in this part of the investigation so no complicated description of appara- tus and equipment will be necessary. Petrographic Examination The samples used had been previously crushed by the Michigan State Highway Department Research Laboratory for the purposes of their investigation. Therefore, a general megascopic description of the ledges was not practical. Several of the larger fragments of each sample, the polished sections prepared for the quantitative microscopic analysis and the chert, sand and clay fractions of the elastic residues *were examined with the aid of a stereemicrescepe.. These descriptions Aeoeeaeeoov .H oases .mmeum Haeoom .oaou one use wsoeoaawu swans .338 622% 8: neonanose nsoaom use I eaaaaepnhuo mamas“ .> menu .oaaahm eomue .eecnfiaom Ioaawm .ocoanEHH human .«Hu =NH enocxoan .mnaona .ueecsoa .vemaeum .nsozomwaammou ova: .nsouemaaweeou .«Hm .omceeL 0 one m Ionsomn phosb ndneoa op 02H on Hump mo been I ocHHHmunhgo hHoGHH .> menu comm uneazamo .eoxooeee 235:8 .vennaaoa use huenowaonocoo .oGOpmoeaH esooomaaauwm homes can mnomxoan .oafinha ovens .eouczou .eocfleuw hone .nsonmpoae kno> .wsoaouaawnnom .«He m emu m .meaHHoonweaepedHo adwwos 0» seem nsouom huo> Hush .ocaHHmpnhuo hHecan .> new comm peeacpdo .emet :28 oaeaoaoe 2.: .eonefiaom .huecoavouocoo .eoxeepam .ecoanesHH edoeoea eon neocxown .ouauha momma .eoeddou .eocwenw end nsououaaamnou Iafiwae hapnmwan .hecmo .oncev m emu b .nouaaoo venoumsao suavea 0» e aneonanoee was I ocHHHeamhpo haecum .> menu comm poeanuso .opfiuhavoomue .naaeaehuo unmade .owpacoosaao .eenmflaom .veensou .mpcosmeau Heuoo .memm .a emeea 0 63.300 vacuums .wmcmmuw ocH assesses 93m 33m one.“ pass. , .ouwahm moans .eoxeeupm .nmenm Hanson .oapficooseao .eozmaaom .nsoneuwawnnou op oaoe masonno «mom .ecoauoaaa =0 enoaxofinj .naepmmuo nausea .eeucsoa .vocwmuw apecowponoaoo amen nzoooeaawmnm muo> .esouoa 9 among m .neuflaoo nonopnaHo adfivea o» scam was mammaaose op“ : .HHm .ocwHHmwmhno hHonHm menu comm useznusmm .eepaoe hanaem .naonqnoee new one .o:oumee«a nsooomaafinme nacho xmq umecxoane .eufiuma comma .eeucson .Uecfimum mnecowueaoeoo heam Imwawmmom .hazono mze> .hdcee o swoon 4 .uauacooseau adficoa on same .msonomaHHneom new ocwaaepmmno haocmw,huo> Mama one» unoznpso .ncwmaea Hence .ep< .oGOpmeeaa nzoeoea .ouauha moans eonnwaoa .nsozauoee Mann use Iaamue econouwaaueou humane zwa onemxowna .naamanmuo sewage .eoucson oa amasmce hue20finouomoo heaw huo> .hecen .«Hn .nsoaoa .«He m omeoq m .nebwaoo nsooowaam nan .eonweaw sewn .ndonomwaweeou ova .ocnaaepnhuo haemam memo comm unassumom .eaceswenu Hence one eouohup oefioeo 8.: .328 SE: .boeofioeo .88 .< .33 m .mamunhao «passe acaeam magnum «elem InOo Adam «swam oaem comm nuuoz [laddmmmHMMImmmmmm‘ .ovcoaman .nnou esoooaawn .o«a«d005eao .ueneaaoa .edoueuwdaonou a .oaoe .eeaIemdm .oeounoswa ama enocxowna .eooaaoo one ueezsou .eoouOn saws: .nunoeweuu snoeueaawume heme» .«Hn canoe 4 among a coma uncanpaom .HmIOHIS eoeoodoo .338; .358 :83. .83.!3 euoaaem .wawaenfiz bond:— vain 333m .228 HO moaunfiuoweo oaqemm woonsom WMMEdz endeeea guano eeufleaee .ucoaaooe eonewaom .eaneawenu oaeweumwe assumes no coapqwuoeoe ewnaeumouaom .eseaeou been one .euoonoua noueeeoa neoane> you .cemfinofia .mcamceg poem mhuopeaonfiq noneoeem aceaaueaon megawam opeam sewage”: on» an eecwwnne :ssHoe man» :a moanecmweee heaped use nonssz .eeneoe eadmxoopn one.mo«u«eoa hunesow . H oases .eepmonu .eaonouaaaoeou m oaqaem .uocaeam onam hno> evens .huecoauonocoo .ecoanesaa owaaaoaoe ocaa when :oaon.MIN ma .ebanma eoen Avewcdon HHow hao> one manganese e» IHevnhuo onneoo op scam me use we: no .eoaawe .nefiow vouoaaeom .ofleda 32.. so? eooeeeee 6:3. Define 68:33 e 383 .ecoz .eecaeam .emeum guano esona caufiaoaoe neoaom hanfien .ocda .uusn zoaon.oa 4H mean ooenp unwaam Inoae neasmce menu IHemehuo aswvoe on new“ ce use so: aoumo .ueaeoum .ecoueesaa m cameem .ecoz .eeauoe Ham: .ocoz owpflsoaoe eeonom hzo> emcee weweenc <4 ma acecfieam emaea hao> bu ecaaaepehno macaw“ me oafimxoovm .Hmma uepoeaaoo .ofino .hucsoo seven .ecopmeeaq enezeaon .Npueeo .mpcosweam .mu«o> venoupeom .epceawenu H mamaem ecaaaeuehao .ecoz esonmuoae epwnw Haeeom ensonno pceecsnd maaveuu mo NH mo eoeue Mo eoeau anwwam .ecoaeoaaa emcee hao> me oawaxoopm .Hmma noveeaaou .cem«50H: .hpnzoo oenwxoe: .ocoueesaq «usam pcusm .neaaeee a mamaem no» e>one mo ovfinoaeoo wadeeuo mo AH oawqxooum .neasmce nee .esouomaaaeeom .ecope on woeceoa .eocaeuw tee esosauoae Hump IoEHH memes haanmaae canoe .0 one 0H oawpha eceecsne Heaven on seam oa even: no ocean on eeaaaeaehao Agenda heme? oven capo .HmIoana eouooaaoo .eemanoaa .hacsoo cones .eaoaneada peachem 328d. 35:— vamm oneseee stone moaamaaonon camaem wooaeom ones: aee.. are compiled in Table 1, pages 25 and 26, and need no further explanation at this time. Their special significance will be touched upon in the conclusion of this report. Removal'pf‘épig Soluble Carbonates. ‘ Approximately 600 grams of each sample was washed to remove all dust, clay and foreign matter, thoroughly dried, and crushed as finely as possible with a steel mortar and pestle to help speed the complete digestion by acid. The relatively large sample was used in order to reduce the proportion of experimental error and to obviate the necessity of performing the minor details of the procedure with several smaller portions of the samples and then averaging the re- sults. The samples were carefully weighed and each placed in the botto- of a very large beaker (£000 ml.) in preparation for acidizing. The finely crushed limestone reacted very violently when treated with dilute hydrochloric acid so the extra large container was necessary to prevent minor losses of clay and other fine material contained in the voluminous rising froth and bubbles. The samples were first wet with water and then the 10 percent solution of hydrochloric acid was slowly added. The treatment with hydrochloric acid was continued, gradually increasing the concentra- tion of the acid, until complete digestion of all the soluble carbon- ates was assured. m Sieving 3:9. Sepgate Clastics - Chert, Sand papg Clap. In the usual insoluble residue study the elastic residue would be _.K.u. .5» PIE «INN .u PIN run nuM PI “FIN. 1“ -.~ \.. COMPOSITE DATA SHEET STANDARD ACCEPTANCE MING W1C“. ANALYSIS _ Percent Total Clastics i Soluble m g g lSoundness ( ub e R Carbonate s o. a w: a: Percent B '3 6 54? g o g Total Silica a Source* 0 g :7 0 lb '4 g vs 2 t: a 5’? {3- e 3" 3" 3’ ‘5‘ o 8 r 3 ° 3 a 3 ag= E . mares: . s . . o c: o on e e e :1 a n P- n e n g g a o s e o c+ ct ('0’ :38 fl 0 O 0 O O H < m m :3 c+ u g o to o n. a a figssg a s r V E r a a: 5' m s: z: a a s 5" 8 B 8 a a. slum Number 1 2 3 II 5 6 7 8 9 10 ll Bayport Limestone, Huron County, Michigan. Collected 10-10-51. 3.1 1.6.80 21.1. llI.3 2.8 0.9 1.9l 11.5 85.7 3.2 2.5 63.1.1 27.1 13.L 5.1 1.1 4.0 8.3 86.6 17.9I 3.3 66.50 5h.3 31.8 20.5 16.9 3'6! 11.3 68.2 26. 1.2- 21.91 13.9 19.6 12.6 6.2 6.1. 7.0i 80.1I 22.5 1.9 73.10 28.9 16.8 1.5 0.6 0.9 15.)! 83.2 3.j 2.60 2.0 70.38 27.8 16.8 21 0.6 1.5 16.7! 83.2 lj Southwest fac 9I 7 ladge E 12.65 1.3 12.35 18.5 8.2 5.0 0.3 l..7 3.2 91.3 2- Thickness_36” Southwest fac 2.63 1.1. 11+.89l 11.5 19.9 15.54 1.2 llI.3I lulu 80.1 16.1 t, . Southwest facd 2.66 0.9 33.51 8.1+ 6.3 3.1 1.0' 2.1 3.2 93-7 S-SW 2.68 0.5 11.27 5.3 II.2 2.1 1.0 1.1 2.1 95-3 0-2 2067 1027 11.60 13.1 116.6 8011 005 7efi 60 85'“ 9°i Table 2. (continued) COMPOSITE DATA SHEET (continued) STANDARD ACCEPTANCE TESTING GEOLOGICAL ANALYSIS :8 Percent Total Clastics 3 Soluble g E a, Soundness (Insoluble Residue) Carbonate m a m \n > to w- Percent ’U f" 4 VI < 5‘ Source* E In E"? 2 2 o 3 Total Silica 3 z t“. P E 333 5“?» 62 3’ 3’ ‘o" c.- '3 5 r 3 3°° 5° 2 a g as: . e ° egress g a e 3%" '1 g D" O ’1 a g 3 fl' CV d' 05' O U 4 a g g 8 I» e e o m 0 ex» [2‘ w- o :I c" n s :r H n. o a 3 figsag" : é a v a z: a s 5' a g o m R. o :3 m a Column Number 1 2 3 A 5 6 7 8 9 10 ll Burnt Bluff Limestone, Mackinac County, Michigan. Collected 1951. Stockpile if 3 7 A 3 12 6B Grading 2.67 0.58 0.57 13.8 2.0 0.02 0.02 none . 1.2 98.8 - Sam le 1 , 7 N 7 laware Limestone, Lucas County, Ohio. Collected 1951. Stockpile 13 LA Grading - - 11.30 2.0 1.7 0.8 none 0.8 0.9 98.3 '- Sample 3 Bottom new cut ll. 10' below surtl 2.51. 3.81. 3.581 3.2 5.1 3.1. 3.1. trace 1.7 9i..9 - Sample A Top new cut 15 2"3'b310‘ 8111‘ 2.53 3.52 707 he? 30“ 2.6 0.3 203 008 96.6 " 1 Sample 5 *Quarry position and stockpile source. Number and letter designati Lona in this colum assigned by the Michigan State Highway Department Research Laboratory, East Lansing, Michigan, for various research projects. Table 2. analysis on limestone aggregates. A compilation of the results from standard acceptance testing and geological removed at this stage, dried, separated and weighed. Because of the very large percentage of clay present in some of the samples diffi- culty in separating the clay and coarser elastic fractions by screen- ing after the residue had been dried was anticipated. Wet screening seemed to be the best method of effecting complete separation of the chert, sand and clay so it was expedient to do so while the residue was still wet and thus eliminate extra and unnecessary steps. After complete acidization of the samples the residue was allowed to settle in the bottom of the beaker and the excess acid solution was decanted with a siphon. More water was added to the remaining solute and residue, at the same time washing down the clay clinging to the sides of the beaker, then the settling and siphoning process was repeated. This cycle was duplicated several times until the solution no longer gave evidence of being acidic when tested with litmus paper. This neutralization of the acid was necessary to prevent damage to the delicate copper wire screen of the 200 mesh sieve used to separate the clay from the coarser clastics. The No. 200 and No. 35 (U.S.Sieve Series) size sieves were used to separate the three elastic fractions. The two 6 inch diameter sieves were nested together and placed over an appropriate size (2000 m1.) beaker. The residue was carefully and thoroughly washed through the 'sieves with light water pressure being careful to remove all clay from the sand and chert. The No. 35 sieve served not only to separate the sand from the clay but also prevented the sharp harsh chert from coming in contact with and damaging the delicate 200 mesh screen. an All that material passing the 200 mesh sieve was considered to be clay for the purposes of this study. All that residue passing through the No. 35 sieve but retained on the No. 200 sieve was con- sidered sand, and all that material larger than the openings in the No. 35 mesh sieve was chert. The No. 35 sieve was selected by trial and error as the sieve size which allowed nearly all the quartz sand grains to pass through but successfully retained the majority of the small angular chert fragments. The chert and sand fractions were oven-dried, weighed and put aside for further examination. The amount of chert and sand were computed as percent of total sample individually, Columns 7 and 8, and combined as total percent silica, Column 6, Table 2, pages 28 and 29. The clay passing the No. 200 sieve was allowed to settle thorou- ghly and the excess water decanted. Due to the large amount of clay and its extremely low permeability it was found that the ordinary lab- oratory glass funnels with filter paper would not be satisfactory for removing all the excess water from the clay fraction. Removing the ‘water with this equipment would have taken an interminable amount of time. A No. A Bfichner funnel, placed in the top of a suction flash ‘with a rubber stopper, and connected to a continous vacuum, worked ‘wery well for this stage of the separation. As the flask was evacua- ted enough pressure was created to drive the water through the imper- meable clay where it then collected in the bottom of the flask. It was .found that if the vacuum was left in operation after all the excess water was removed that the flow of air through the clay tended to dry it out sufficiently so that the thick clay residue shrank away from the sides of the funnel for easy removal and was solidly compacted for ease in handling. The clay "billets" thus produced were dried, weighed and their percentage of total sample computed. The results are found in Col- umn 9, Table 2, page 28. The amounts of each of the elastic frac- tions, chert, sand, and clay were added together and deducted from the original weight of each sample, the remainder being the percentage of total soluble carbonates. The total percent carbonates is tabulated in Column 10, Table 2. Separating the water from the clay would probably have been quick- er and easier by evaporating in an oven. However, alteration of the clay by inadvertent burning or baking would have been detrimental to further investigation of the clay fraction. The more tedious mechanical separation procedure was selected in order to avoid any adverse effects on the clay minerals. Qpantitative Microscopic Determination of Calcium.!§gpesium.Carbonate. Exacting separation and determination of the proportion of calcium magnesium carbonate (dolomite) in combination with calcium carbonate (calcite, or limestone) is a difficult quantitative chemical procedure involving considerable skill, equipment and time. It was not felt that such a quantitative chemical analysis was in keeping with the scope of this study. However, it was desireable to determine the extent of dolomitization, or percent dolomite, inasmuch as porosity is probably one of the major factors in the response of a limestone aggregate to the various soundness tests. It is becoming a generally accepted fact that increased dolomitization is a direct cause of increased porosity in limestones. The following method using ordinary.mineralogic staining tech- niques and a mechanical integrating stage in conjunction with a polar- izing.microscope was suggested as a substitute for determining the degree of dolomitization of the samples. Calcite and dolomite are closely allied minerals which are diff- icult to tell apart even under the microscope. A number of methods .. employing various stains have been recommended for use in differen- tiating between the two minerals. Some stains are considered to be better than others by each investigator but all seem to depend on the greater solubility of calcite as the distinguishing feature. In preparation for the quantitative microscopic analysis three to five fragments of each sample were either cut-off with a diamond saw or ground off flat on one side and polished smooth with a hori- zontal lap using emery and aluminum oxide abrasives. These polished sections, if the opposite side was left uneven, were mounted on glass slides with a small amount of modelling clay. Before mounting the polished sections on the glass slides they "were first stained using the Fairbank's Method, a refinement of the long used Lemberg's solution. The staining solution was prepared by bringing to boil: 0.2L grams haematoxylin (legwood) 1.6 grams aluminum.chloride 2A cc. water The solution was allowed to cool and a small amount of hydrogen peroxide added to completely oxidize the haematoxylin to hematein to insure uniform results (Fairbanks, pp. 126-127). The prepared polished sections were marked for identification, immersed in the Fairbanks solution and then boiled for approximately 5 minutes. This boiling is not a previously recommended practice but was necessary to insure distinct staining of the calcite to aid in the planimetric measurement which was to follow. Several other staining methods were attempted including the Copper Nitrate Method, the Silver Chromate Method, the Potassium Ferricyanide Method and the old Lemberg Method (See: L.W. LeRoy, "Stain Analysis", Subsurface Geologic Methods, 2nd Ed., Colorado School of Mines, Golden Colorado, 1950, pp. 195-196, for a complete resume of calcite-dolomite staining methods). With practice it was found that the Fairbank's Method produced the most consistent and most distinct results. The stained polished sections mounted on glass slides were then placed in the E. Leitz Integrating Stage under a polarizing .microscope. The integrating stage is a precision mechanical device for measuring microscopic planimetric distances on thin or polished sections. The apparatus has six independent measuring spindles ‘ for measuring proportional amounts in one direction, plus one spindle on the side for offsetting the line of measurement to re- turn.on a new traverse. In this way a planimetric determination of the proportion of unstained dolomite and stained calcite was accomplished. To reduce the proportion of possible error in measurement, the total length measured should exceednat least 1000 times the average size of the individual crystals. The unstained dolomite crystals averaged 0.5 mm. so the total planimetric measurement exceeded 500 mm. for each polished section. The percent of dolomite in the total sample was computed from.the total carbonate fraction by first cal- culating the average percent of dolomite from the total distances measured on each of the three to five polished sections for each sample. This total percent dolomite (unstained) is recorded in Column 11, Table 2, pages 28 and 29. Although this planimetric method of determining the percentage of dolomite in a sample of limestone may need qualitative chemical confirmation to become generally accepted, the range of percentages determined by this method do approximate and include the percent of magnesium carbonate in a composite sample of the Bayport limestone determined by chemical analysis by the Michigan State Highway Depart- ment Research Laboratory. Chemical Analysis Bayport Composite Burnt Bluff Composite 3102 15.5h$ 0.83% 3203* LW 0.75% CaCO3 75.83% 88.28% The Burnt Bluff samples from.Mackinac County and the Delaware samples from Silica, Ohio were not included in this quantitative micro- scopic analysis. Even though the Burnt Bluff limestone is almost '* Tri valent iron and aluminum oxides. 10 percent magnesium carbonate by chemical analysis, the samples were so dense that none of the staining methods attempted were successful in penetrating the calcite. The Delaware samples, because of their unusually high absorption, reacted just the opposite. The whole sample became darkly stained even without boiling so no differen- tiation between dolomite and calcite could be discerned. Fragments of the Delaware samples reacted comparatively slowly when treated with dilute hydrochloric acid and are probably highly dolomitic. C151 Mineral Stain Analysis Clay.minerals of the montmorillonite and illite groups expand and contract significantly with wetting and drying (Blanks, p. 1.09). The expansion and volume change resulting from the hydration of mont- morillonite type clays is accompanied by potential pressures of several thousand pounds per square inch which may greatly exceed the tensile strength of concrete (Rhoades and Mielenz 1948, p. 35). The expansion of a stratified limestone may differ in different directions and can expand more than 0.1 percent in length upon wetting (Rhoades and Mielenz 19h8, p. 35). Even though some lime- stones may be considered sound according to standard acceptance tests they can still cause deterioration of concrete when their ‘interstitial clay swells as it absorbs water from the concrete (Blanks, p. h09). Identification of the various clay mineral groups and especially leachrspecies within the groups is extremely difficult due to their minute size. Optical methods are very unsatisfactory and uncertain. Chemical, Xpray, electron-diffraction and differential thermal-dehydra— tion methods are required for precise and reliable identification. Several stain tests have been employed to aid in the determina- tion of the various clay groups. Caution should be used in the applica- tion and interpretation of these tests because impurities and complex clay-mineral associations may lead to inconsistent and erroneous results. (LeRoy 1950, p. 197). In order to determine if the excessive clay present in the Bayport limestone, occuring both interstitially and in concentrated clay streaks, was of the swelling (montmorillonite) type the following clay mineral stain tests were employed. Inasmuch as the clay residue had already been thoroughly acidized, washed, sieved and dehydrated, the malachite-green and crystal-violet tests were used. By using both tests the results were mutually confirmed and should be more reliable. Both the malachite-green and crystal-violet tests make use of the same technique, the only difference being the organic dye employed. Both dyes were prepared by dissolving 0.1 gram of crystal-violet or 0.1 gram of malachite-green respectively in 25 cc. of nitrobenzene. Approximately one milligram of the acidized clay fraction placed in a watch glass and treated with a few drops of either dye solution reacted by assuming characteristic colors as given in Table 3 below. SUMMARY OF CLAY-STAIN RESULTS* Mineral Group I Malachite-green Crystal-violet Kaolinite IGreen Violet ’ Mentmorillonite Yellow to greenish- Yellow, greenish- Yellow yellow or orange yellow Illite allow Bark Green *after L. W. LeRoy, 1950, Table 7, p. 199. Table 3. Summary of clayemineral-group stain results using acidized samples. The stained clays were observed with a stereomicroscope in re- flected light using fairly high magnification. The results of the clay mineral stain tests on the aggregate samples are tabulated in Table A, pages 39 and LO. Statistical Compgrison of Physical Testing and Geologic Analysis. Upon completion of all the tests and analysis Just described we find that we have a mass of random data. From the three engineering tests we have arrived at various percentage measures of soundness; from the geological analysis we have isolated six percentage measures of the basic constituents of the samples which we hope to relate to the measures of soundness. These together times the 15 samples used for the s tudy give us roughly 135 individual quantitative measurements (of the variables on which the concrete making properties of the sam- ples depend. Individually these percentage figures have only a very limited value and then only in a general sense. It is essential that we group 4‘ AeoscaaaoOV .s oases .eawcfiaoex m magnum .ovficfiaoex new .coou mo momma no“: .poaoap mo momma econ» mcdummc mo HA oaficoflaauoEvcoa com zoaaom nomcaou .ouwcoHHHooSo:04 so“: :oHHom Swansea eafioxoowm .ouacwaomx no woman saws .poaofib mo moon» .0 owned 0H .oaficoaaamoEuco: .zoaaoh cmmcson .opwcoaawooaumo yup“: soaaom newsman oomu swamp .ouamaaoex :NH unomxoflna mo comma no“: .aoao«> no some» spa; c omooq m .opdcoaawmoaamoz .soaaom seamen .euwcoHHHnoaamo newcoemm o» :OHHoh xmmn ewmwxvuoznvsmm gsm unocxowna .zoaaoh m omuoa m .opwcoaaamosoco nsfiaooom on :OHHe .opacoaamuoapco .voHoH> use :OHHo oomu muosnosom gem unooxowna .zoaaoh .zoaaeh m omuon b .ouficOHHfimoEvmo newcoomm o» xOHHo .ovwcoaaamoepcoa nnwmeemm use onHo oomu neomnummm .ouwmaaomx .ooao«> some» new: ovaaaa we come» one coemm .o emoog o .mchOHHHuoeoco: .aoaaom nnwmzommrl.ovmcoHHHuoanco .30HHoh newcoomm 3mm oomn neon. .eawmaaomx .poaoar ab unocxowna .osacamoax one «o oases as“; mongoose smegma as“; a omeoa n ouwaoaaamoaaaoz .cooam one :OHHo .opa:oa~auo&uco newcoomm op soaaoh mama oomu uneszammm. :Ns eoooxoana .opamdaomx one .epacwaomx one o emqu 4 oaH:0Hawooapco: .coomm one seaflomfl opH:OHHmmoEooo= .poHow> use moaflmr one no sow .opaoaaomx .aoaoa» :NH unocxowzh .osaaaaoaa use no ovens and: asaaosss smegma noes; m omeon m ouadoaaamoéamo: .moeum one roaaoh some .ouwcoaaflmoauco newcoomw op zoaHo» amen econ umotnvpmm. .eaficaaomx no .mo .uoaoa» yo oven» oaacoaaamoaaco: .coomm was aoaawm commoadfimoaaco: :oaaoh nnaaoomw .soHHe comm npmoz .eaacaaoex .opacflaoex .peao«> :ma noocxoanh mo comma no“: .noacooum we come» and: anamoama pnecam snarl < owuoq H .opficoaafiaoEano: we soon» and: .moaaew .ouwcoaaamoaoco: anamoonm ou sedans xnen ooeu onoszvsom .HnuoHooa vovooaaoo .cmmwnowa .hpcsoo moms: .ecoonoeaq gmomhom depends epasoom madam Heoocda npaaoom camaw . _ menmsz _ ammo sausages. noose cameoaaas , some easemeeeosm Jasao ooHomleeouaao mooazam oaaeam mamas 2H use :0: mos .opwmoaawuos 4 magnum tamoE mo oomup .umSn moan: .oa 4a new; .ovficwaomm .onHoh meme use cocoa .ouameom .uoHoa> poo 30c souvom m omoasm mcmemac <4 mm .oowcaaomm .Coomu .oawmwaomx .peao«> oaamxooam .Hmma vouooaaoo .oano .mamdoo omega .ocopmmeaq ommzwaoa .oaanoaawmoa .ouflcoaaamoa H oaasmm 15” .3 82s .03?“ -20”. so 82a £33m sass wages as 3 no“: unecaaosx mo woman rue: owmmn new: oaficfiaomx go some» new; poaoa oaaqxoopm .Hmma cocooaaoo .cmmfinoaz .mpcsoo oeaaxom: .ocoamoadq madam use fiancee: smao‘acaaemovoaa mvadmom a “mom ”magnum eassm some: n amused: asHo cocoa onwnoeaemelmeHo onscwsoeomm cameo acondo oossaoaoov momma 2H Hmpmzmo MZH: w4 8 10 12 1h 16 18 20 22 2h 26 I 8 Percent Sand 0. igurg 5. Comparative relation of total sand to magnesium sulphate soundness, freezing nd thawing soundness and absorption. g p7 o 1 o 3 8 317-23 .96 5 vcslh. 984-0. 67x 7 / . o r -o.1732 m V1139 /’ E70 o 7:570 “'“r .O'L’615 /’ o o .c g 60 4 E60 l/ a 3* ° 350 / 9? 5° ’4 5 o / k‘ ‘ § |‘(/, .3 /’ Z a” 7 s“ ’ ’ J a o D 2 “" 2S - no Sim-— :30 a. / T‘s—vi O I) .§ .3 0 £20 13,20 3 I/ 0 O I 2 o I / :5 ° 0 ’/ 631° I , 1° ° A/ u ,1/ w I I 8 Percent Unstained Carbonate X ' Percent Unstained Carbonate a. b‘ Regression Line Tc 3 1. 511-.sz C 0 8 a Standard Deviation = 0.905 3 Correlation Coefficient r 8 0.3507 0 ‘2 43 C 0 0 Id of II 2 H 0 2 l. 6 8 10 12 11+ 16 18 20 22 2h 26 I 8 Percent Unstained Carbonate c. Figure 6. Comparative relation of total unstained carbonate (dolomite) to magnesium sulphate soundness, freezing and thawing soundness and absorption. 0 3 r’ l L :3 #f // €30 Yc=25.76+2.121 _ :30 A g sy-zms " I] I I / m o r I0.16I.6 .5 [29' / o 1 37° 0 E” / 7' 5. o 5 3,60 —— 560 l I “0 I 3 5 / :50 /l :50 I c: o / a , I 5 “‘ / / ' 1.0 / 31.0 j 5,: __/_ / 5‘ ° K m / I “\30 ~30 to g V f l n 320 ° 320 ° / Ic=z..66+_2_._;§x_ #3 / a, I sy=905L § 0 . £3 0 r =o.69o7 M1.0 a 010 v l I ‘3 °I MTV 2‘ F 1' I 1 u ‘b . H 0| l/ l *‘ _ / l I 0 10 20 30 1.0 50 0 10 20 30 1.0 50 X 8 Percent Total Silica X = Percent Total Silica b. 8. 5 Regression Line Ic = l. 3 Standard Deviation a; a . 2' Correlation Coefficient r 0.2 E «t ‘5 8 h of u >4 0 2 l. 8 10 12 ll. 16 18 20 22 21. 26 I = Percent Total Silica c. Figure 7. Comparative relation of total silica (sand and chert) to magnesium sulphate soundness, freezing and thawing soundness and absorption. § / X a: I I +3 I 380 / é" so .[ l / ,3 /l .5 I c=l.37+l ,211 / / +370 I C 70F? =0.8123 / e /' -H I ‘11 ° 3 / o ,2 7 // / 3» / / 5 / 1' ‘° ’ ° 1 m o I 050 o 50 :3 a /7 I/. ;: I A/if/ /’ ow 23y / r3“) / /5’Sy a / / a / 1’ (:3. 0 I n - / In30 1 N30 9 u’ ,7 o) l / g / O / E O V, 3 ' 20 +920 -- a / r: 0 I g 1 1,96. 064-2 2,9111 3 O o I 2 O o /, 3,919 73 3 EP, 0210 I 0! ”.6ldi5 “‘10 p I u a} I I I I w / I I | *‘ J I 0 10 20 30 1.0 50 0 10 20 30 £0 50 X 3 Percent Total Clastics x I Percent Total Clastics a. b. 10 8 8 egression Line Yc ”122 3 Standard Deviation SI = 3 Correlation Coefficient r - 0; o '1 6 Q *5 3 h u 93 ll 2 H O 10 12 lb 16 18 20 X = Percent Total Clastics C. Figure 8. Comparative relation of total clastics (chert, sand and clay) to magnesium sulphate soundness, freezing and thawing soundness and absorption. : EL 7! .. 5 *f 1 s 580 . I 9: (3 5 9% I I 570 3 11 _°_ 33 s a ° 5 60 . I 5 4% .5? U . I I: 050 e I 3 o a 9: ~ 3. IT 0A0 I o E «'3’ / 1 O D in l “‘30 (2 30 E a 3.; ’ .3 I ' "2°L I 22° g I Ic=18.73+é_-_&3X — 8 1310.676;ij __ 3410 1 3,325.19 £ 0 1 12038 ‘3 o 1‘ '0-2646 "1 r 30.8167 '— I! n I 1 l l ..V° |[|l| “l“ll Ill 0 10 20 30 w 50 0 10 20 30 1.0 50 X 8 Percent Absorption I = Percent Absorption a. b. to = Regression Line 8 Standard Deviation r = Correlation Coefficient Figure 9. Comparative relation of percent absorption to magnesium sulphate soundness and freezing and thawing soundness. 53 llull‘flk’lil’i very high coefficient of correlation; 0.9102 for magnesium.sulphate soundness, and 0.7655 for freezing and thawing soundness. These are considerably higher than the correlation coefficients of the other constituents of the samples to soundness. (2) How well the points fit the line of regression which produces a very narrow band encom- passed by two standard deviations. (3) The high degree of dependence of soundness to clay which is illustrated by the factor 'b' in the estimating equation and visually noticeable by the steepness of the line of regression, which almost approaches vertical for magnesium sulphate soundness. In other words, for each one percent increase in clay we can eXpect almost 5 percent (actually A.85%) increase in average loss on 5 cycles of magnesium.sulphate soundness, and a 2.08 percent increase after 25 cycles of freezing and thawing in.water. (A) The closeness of the point of origin of the line of regression to the zero point is a definite indication that only 0.hh percent of the loss from 5 cycles of magnesium.sulphate soundness is due to reasons other than clay, and all but h.2 percent of the loss on 25 cycles of freezing and thawing is due to clay. Attention should be called to how well the points line up Just below and parallel to the regression line in Fig. 3b. A few stray data draw the line away from the row of points but their remarkable alignment attests to the closeness of correlation between clay and freezing and thawing soundness. Such a close alignment is more than 'we could reasonably expect from such a small statistical sample but is so close that we could almost accept the factor 2.08 as the math- emetical coefficient defining the effect of clay on the freezing and K). thawing of limestone aggregate. Such a “clay coefficient” firmly established by the study of many more samples would be invaluable in estimating the soundness of concrete aggregates. It should be noted that those samples drawing the regression line away from the concentrated row of points possess a high percentage of chert or are heavily streaked. Two special characteristics of the clay prdbably aided in its predominating influence on soundness. The high percentage of clay in the Bayport samples was prevalently of the swelling types, mont- morillonite and illite (Table A). The Burnt Bluff and Delaware samples had only traces of montmorillonite. Nest of the clay in the Bayport limestone was concentrated in thin scattered shale streaks (Fig. 10b). The individual undulating streaks do not persist more than a few inches but are concentrated in certain ledges in the quarry forming heavily streaked lenticular zones which may continue for several feet. When the rock is crushed into fragments less than 1 inch in size these shale streaks produce zones of weakness transacting the individual fragments. Water entering these porous shaly zones by capillarity is probably the major cause of disruption of the samples. {The progressive entrance of water and the pressures produced by the swelling clay is in itself force enough to cause failure. Add to ‘this the eXpansive force of freezing water and there is little doubt that the Bayport limestone is inclined to be unsound. Such jligmtly streaked to heavily streaked argillaceous limestone com- prises from 30 to 60 percent of the Bayport limestone section in Huron County. The Influence 93 M on Soundness Chert has often been blamed for being the major cause of the disruption of concrete which has been subjected to freezing. This may be true when chert is actually encased in cement but the results of this investigation show a relatively low correlation between chert and soundness. At first glance (Fig. La and b) there appears to be a reasonably fair correlation especially between freezing and thawing and percent chert. However, a closer observation shows that almost all the points are concentrated on the lower end of the chert scale. This illustrates the fact that we do not have a very representative sample for studying the effects of chert to soundness. With ninety percent of the samples having less than 3.5 percent chert we cannot rely on the statistical measures which we have computed. The effect of chert on the soundness of the samples has been almost entirely masked by the predominance of the effect of clay. This is shown by the high value of 'a' in the estimating equation. There is 25.hh percent less with magnesium sulphate soundness and 12.12 percent less from freezing and thawing due to other reasons, probably predomdnately clay. A much higher correlation between chert and soundness would be expected from an examination of the chert residue. The chert fraction Figure 10a. Photomicrograph of aphanitic fossil— iferous chert and amorphous chert fragments 6X9 tracted from sample No. 3 using dilute hydrochloric acid. Magnified approximately 5 times. Figure 10b. Photomicrograph of finely crystalline argillaceous limestone showing the undulating con- centrated clay streaks. Polished section using reflected light. Magnified approx. 5 times. Beaded 31w. H9 E. Grand River WIlrIamsmn. Mich. Pram or use consists largely of non crystalline silicified fossil remains, Imostly corals and bryozoa, whose acidized remains are liberally penetrated with voids (Fig. 10a). Of course these voids are all filled with calcite in the natural state and when the limestone is used as concrete aggregate. But much an impure intimate combination of calcite filled voids in chert with their differential thermal ex- pansions should be very conducive to unsoundness when subjected to freezing, and thawing. Freezing of a calcite filled coral fragment similiar to the one shown in Figure 10a should open up countless expansion cracks which would make way for further penetration of water and progressive disruption. The failure to achieve a dependable correlation between chert and freezing and thawing soundness in this investigation should not discourage further attempts along this line. The closeness of the points to the regression line (Fig. Lb), even though they are all concentrated at one end, is indicative of what might be accomplished using a good representative set of samples with a wider range of per- centages of chert and without the detracting influence of clay. The inferiority of the results obtained with magnesium sulphate soundness compared to freezing and thawing soundness is of consider- able interest. It suggests that the pores and cracks in the chert, or produced by the chert, may be so minute that the magnesium sulphate solution was not able to penetrate as easily as water alone. It also suggests that at least part of the failure caused by 25 cycles of freezing and thawing may be due to differential thermal eXpansion. The Influence g_f_ M 25; Soundness. Although there was considerable sand in some of the samples it has produced little or negative effect on the soundness of the aggre- gates. There is only a minutely positive correlation between the per- centage of sand and freezing and thawing in water (Fig. 5b). This correlation is of little consequence because of the widely scattered array of the points around the lower end of the line of regression. The influence of sand to 5 cycles of magnesium sulphate soundness pro- duced a very negative correlation (Fig. 5a) which is of doubtful value. The sand occuring in the Bayport limestone is thoroughly dissemina- ted throughout the rock. If it were concentrated in thin streaks then it might possibly be a source of trouble. The sand extracted from the Bayport limestone consists of poorly sorted, fine to coarse grained sub angular to sub rounded quartz. The sand from the Delaware samples ‘was both very fine and very coarse grained well sorted well rounded frosted quartz. Probably any cracks or weaknesses in the quartz grains were opened up long ago while the grains were in the process of erosion, ‘weathering and transportation to their present state and environment. The Influence 2;,Dolomite 25 Soundness The degree of dolomitization as determined by the staining and quantitative microscopic analysis apparently has had little effect on the soundness of the samples. However, this lack of close correlation between dolomite and soundness may be due partially to the method em- ployed in determing the percentage of dolomite but is primarily due to the overwhelming masking influence of the clay. There is a fair correlative relationship between percent unstained carbonate and freez- ing and thawing in water (Fig. 6b). The range of the percent dolomite is amply representative but the amount of deviation, low degree of correlation and high point of origin of the line of regression cast doubt upon the reliability of the relationship. The failure of the method employed to produce a good correlation does not necessarily condemn the technique. The relation of percentage dolomitization to soundness should be firmly established by quantitative chemical analysis. If there is a close relationship between these two factors then staining and quantitative microscopic analysis can cer- tainly be recommended as a rapid technique for determdning the percent of dolomite. The small statistical sample does not give the technique an adequately fair trail. The Influence of Total Silica on Soundness The percent sand and percent chert have been statistically analysed together as total silica to see if their combined effect produced a closer correlation than either of the two separately. This was done primarily to find out if the extra separation of sand from chert was necessary to ’achieve conclusive results. The resulting statistical relation of total percent silica (Fig. 7) nearly parallels the relation of percent chert to soundness (Fig. A) but gives a slightly poorer correlation. This is what should have been expected. When combining the low or negative influence of the sand to the high positive effect of chert the two together would average out somewhere in between. The small degree of improvement in the relation of chert alone over total silica when correlated with the more reliable freezing and thawing test would not warrant separating the sand from the chert in evaluating a concrete aggregate. The Influence of Tgtal Clastics gg Soundness The total percent of insoluble residue, chert, sand and clay, have also been combined for statistical correlation for the reasons mentioned i above. The resulting statistical comparison (Fig. 8) is as good or better than relationships achieved with either chert, sand or total silica alone when correlated with soundness. This is especially true when correlating with the accelerated magnesium sulphate soundness test. This improve- ment is due to the dominating influence of the high percentage of clay when it is added to the silica. The total insoluble residue would be satisfactory for a preliminary estimation of the expected soundness of an aggregate but the clay and silica or chert fractions should be isolated and their individual in- fluence analysed for a thorough evaluation of the concrete making prop- erties of an aggregate material. 211313 Influence 9f the Constituents of Limestone 29 Absorption Each of the constituents of the limestone samples have been correla- ted with the percent absorption. (Figs. 3c to 8c). If the percent absorption is a reliable indication of the expected results from soundness tests then that constituent possessing the highest degree of correla- tion with soundness will also have a high rate of correlation with absorption. Unfortunately, this was not found to be significantly evident. In relating the various constituents of the aggregates, alone and in combination, with the percent absorption, the six resulting correlations were so remarkably similar in almost all respects that discussing the relationships individually would be very repetitious. The reason for this lack of character in correlating with absorption is the very narrow range of percentages of absorption for the samples analysed. This duplicity of the statistical relations of the constituents to their absorption would indicate that the individual constituents have no important effect on the absorption of the limestone. The negative relation between sand and absorption (Fig. 5c) would add to this supposition as we can be reasonably certain that the quartz sand grains have no porosity. This leads us to a very interesting conclusion. Evidently the porosity of limestone is not due to the porosity of its integral mineral constituents but is produced by other factors, probably erosion, leaching and the negative space relationship produced by min- eral alteration. The Influence 2; Absorption 2g Soundness The unexpected lack of a positive relationship between any of the constituents of limestone and absorption prompts an investigation of I; - u‘r— the relation ob absorption to soundness. The comparative relations of percent absorption to average percent loss magnesium sulphate sound- ness and freezing and thawing soundness is shown in Figures 9a and b. A close and very high degree of correlation is found between ab— sorption and the two measures of soundness. However, the narrow range of the data concentrated below A percent absorption does not produce a fair estimation of the relationship between the two variables. Here again, the high percentage failure of most of the samples is primarily [ due to the excessive amount of montmorillonite concentrated in zones of weakness. The effect of porosity on unsoundness has been obscured by this excessive clay. This fact is further emphasized by the high point of origin of the line of regression. The two statistical rela- tions indicate that from 10.67 percent to 18.73 percent of the failure is due to reasons other than absorption for sulphate soundness and freez- ing and thawing soundness respectively. In addition, the Delaware sam- ples, which had very low percentages of non swelling clay but rather a high rate of absorption, proved to be Very sound aggregate material. A Summary of the Influence if. the Impurities in the Bayport Limestone 9_n_ It}; Sgundness In summary, all of the impurities found in the Bayport limestone, 'the clay, chert, dolomite and sand have contributed to its unsoundness ix: the order named. The clay, its type and mode of occurance is the major cause in the failure of the aggregate to resist the simulated forces of weathering. The chert and dolomite probably exert a greater influence on the durability of limestone than was determined by this investigation but their effect was obscured by the preponderant in- fluence of the clay. The sand had little or no effect on the soundness. A few minor mineral constituents, namely pyrite, crystalline quarts and glauconite were present in small amounts in a few of the samples but their isolated occurence was too insignificant to warrant analysis. One constituent, calcite, has been supposedly ommitted from the I previous statistical analysis and discussion. If each of the analysed , ! impurities contribute to the lack of soundness of the limestone, then E._s-.__ _ it follows that the more pure the limestone the more sound it should be. This relation of calcite to soundness may be seen if we examine Figure 8 and mentally add a percentage calcite scale along the hori- zontal axis in reverse (from right to left) to the total insoluble residue scale. Thus, we have a negative correlation with essentially the same values as for total insoluble residue for the estimating equation, standard deviation and coefficient of correlation only the sign of the factors 'b' and 'r' will be negative. The porosity (percent absorption) of the Bayport limestone is secondary to clay as a cause of failure. The lack of a representative number of samples from the Burnt Bluff and Delaware formations prevent us from.making any definite con- I clusive statements about these particular aggregate materials. How- ever, the samples from these two sources are very free of impurities so they should make sound concrete aggregate. The Suitability of the Bgzportggigestone Ior;§gg_gg Concrete Aggregate. Before this geologic analysis was begun it was already known that only one of the ten Bayport ledge samples successfully passed the mag- nesium sulphate soundness test and only one additional sample could pass specifications after the freezing and thawing in water tests. The un- sound samples comprise 94 percent of the section tested. In addition, the thin black brittle carbonaceous shale stringer, commonly under 1 inch thick, which persistently underlies the limestone and forms the base of the quarry has not been included in this geologic analysis. Its clay was, however, also found to be montmorillonite. Any of this shale mat- erial inadvertently picked up and added to the limestone will notice- ably detract from the soundness of the aggregate. Inasmuch as the main cause of failure was found to be excessive montmorillonitic clay any condition in which the Bayport limestone aggre- gate might be used should avoid exorbitant moisture and water. Such con- ditions multiplied by repeated.freezing and thawing should be exceedingly conducive to failure. Avoiding these moist conditions would nearly can- cel the use of Bayport limestone for almost all Portland Cement concrete uses. However, if the dry aggregate fragments were thoroughly encased in asphalt cement the Bayport limestone should be satisfactory for asphalt base, binder and leveling courses in bituminous concrete paving. The Value of antitative Geolo ic sis of Limestone Aggregates Using the 12 percent specification allowable for percentage fail- ing the soundness tests we can now compute the maximum allowable percent- age of clay this geologic study and statistical analysis has deter- mined to be within the limits of sound aggregate material. Reading directly from the regression line (Figs. 3a and b) or substituting 12 percent in the estimating equations we find that there should be less than 2.h percent clay present in the samples in order for the material to successfully pass the sulphate soundness specifications. l“ Freezing and thawing soundness allows almost one percent.more clay, ' or 3.3 percent. This can also be done for chert. man: To be able to furnish reliable information of this nature would be of inestimable value in predicting the quality of an aggregate and in providing an explanation of the results obtained from performance and soundness tests. Of course, the above factors are based on the results obtained from only fifteen samples. They are therefore of very limited value. This same type of analysis performed on many more aggregate samples with a full range of the variables involved would provide us with a means of predicting the performance of lime- stone containing all possible percentages and combinations of impuri- tiOSe Recommendations for Further4§tggy As a result of this investigation many problems which need further study are suggested. One has already been discussed immediately above. This same type of analysis, in fact the data obtained from this geological analysis, should be correlated with various strength tests or performance tests to determine what effect the various impurities have on the strength factors of the aggregates and concrete. The large variety of percentages of montmorillonitic clay in the Bayport samples would provide an interesting study of the relation of swell- ing clay on the coefficient of expansion of the aggregate. To determine the true relation of the montmorillonitic clay in the Bayport samples to the forces of moist weathering conditions, the samples should be subjected to a test similar to the magnesium sulphate soundness test using water alone. Several cycles of soaking in water and subsequent drying should indicate whether the eXpansive pressures of montmorillonite were sufficient to cause disruption of the limestone aggregate. The influence of differential thermal expansion on a cherty lime- stone should be defined. If the dried samples were subjected to re- peated cycles of arid freezing and thawing conditions, the relation of the differential thermal expansion of chert to limestone could be determined. Sources of Error In further quantitative geologic investigations of this type cer- tain practices should be avoided. Closer correlations probably could have been obtained if the same identical samples used in the soundness testing had also been used in geologic analysis. Care was exercised in selecting representative portions of the field sample for both parts of the study but the use of two separate samples introduce minor experimental error in exacting quantitative laboratory analysis which could be easily avoided. In evaluating an aggregate for possible use in concrete it is necessary to secure a representative sample of a whole ledge or quarry face. However, for purely laboratory investigations of the influences and properties of an aggregate material the sample should.be confined to a L f single uniform sample of rock. Instead of trying to include a repre- E a sentation of a whole ledge or quarry face, thereby compounding the i'd variables involved, one uniform piece of rock would be preferable for laboratory use. In preparing the limestone for acidisation in this investigation the samples were crushed with a mortar and pestle. This practice should be avoided. A few of the fine chert fragments produced in crushing contaminated the sand fraction and are difficult, if not almost impossible, to remove conveniently. East Lansing, Michigan Irvin Verne Kuehner 1956 BIBLIOGRAPHY American Association of State Highway Officials, "Standard Specifica- tions for Highwgy Materials and Methods of Sn leigg and Testigg," Part II, -Published by the Association, l9h2 American Society for Testing Materials, "l222_ Book of Standards," Part 3, Cement, Ceramics, Thermal Insulation, Road Materials, Waterproofing, Soils, Published by the Society, 1956. John H. Bateman, "Materials 21 Constructiog, Putman Publishing Corp., New York, 1950. R. F. Blanks, "Modern Concepts Applied to Concrete Aggregate", Trans. Amer. Soc. Civil Eng., vol. 115, 1950, pp. LOB-LBI. Frederick E. Croxton and Dudley J. Cowden, "Applied General Statistics," Prentice - Hall, Inc., New York, 1939. Ernest E. Fairbanks, "A Modification of Lemberg's Staining Method,” Amer. Mineralogist, vol. 10, pp. 126-127, 1925. L. W. LeRey, "Stain Analysis," Subsurface Geologic Methods, Ed. by L. W. LeRoy, 2nd Ed., Colorado School of Mines, Golden, Colorado, 1950. Katharine Mather and Bryant Mather, "Method of Petrographic Examination of Aggregates for Concrete," Amer. Soc. for Testing Materials, Proceedings, Vol. 50, 1950, pp. 1288-1312. Duncan McConnell, Richard C. Mielenz, William Y. Holland, and Kenneth T. Greene, ”Petrology of Concrete Affected by Cement - Aggregate Reaction, "Application_ of Geology_ to Eggineering Practice," Berkey Volume. Sidney Paige, Chairman; The Geol. Soc. of Amer., 1950, pp. 225-250. Richard C. Mielenz, "Petrographic Examination of Concrete Aggregates", Bull. Geol. Soc. Amer., vol. 57, April l9h6, pp. 309-318. Michigan Geological Survey, "Mineral Resources of Michigm ”225 State of Michigan, Department of Conservation, Geological Survey Division. Michigan State Highway Department, “Standard Specifications £25 Roads 39g Bridges, 1950." Published by the Department, 1950. Roger Rhoades, "Influence of Sedimentation on Concrete Aggregates," Applied Sedimentation, edited by Parker D. Trask, John Wiley and Sons, Inc o , 19 50 9 PP. h37'h63 o Roger Rhoades and Richard C. Mielenz, "Petrographic and Mineralogic Characteristics of Aggregates, "mmsium 29 Mineral Agggegates w,” Amer. Assoc. for Testing Materials, Spec. Tech. Publ. No. 83, pp. 20- #8. Harold S. Swat and K. B. Woods, "A Study of Chert as a Deleterious Constituent in Aggregates", Highway Research Bull. No. 9, Purdue uniVo 191020 70 ,_ .__‘__. .— x-a .;~;-, 5:941 , (35E 0:131! a. g ..,, - " Date" Due” Oct 30 '58 Demco-293 A‘— -_.. _