STAFF SUMMARY SHEET "" TO ACTON SGNATURE (Surname), GRADE AND DATE :: TO ACTON SGNATURE (Surname), GRADE AND DATE USAFA/ Coord.{tr lv "" 1 6 2 DFCE... A..,. 1111 USAFA/ DFER Sgn C:.dl A:>.-).l. 23 \,; rs- 7 3 8 4 9 5 10 SURNAME OF ACTON OFFCER AND GRADE SYMBOL PHONE TYPSTS SUSPENSE DATE Name: 719-358-7579 NTALS Henry, Professor DFCE kh 20150301 SUBJECT DATE Clearance for Materal for Publc Release USAFA-DF-PA- CJ08 20150120 SUMMARY. PURPOSE. To provde securty and polcy revew on the document at Tab 1 pror to release to the publc. 2. BACKGROUND. Authors: Karen Henry, Hed Hunter, Ben Fonte Ttle: THE NFLUENCE OF PROCESSNG SOL WTH A COFFEE GRNDER ON SOL CLASSFCATON Crcle one: Abstract Tech Report Journal Artcle Speech Paper Presentaton Poster Book Other:_journal techncal note (shorter than an artcle) Descrpton: For submsson for consderaton for publcaton n the ASTM Journal of Geotechncal TEstng 3. DSCUSSON. NA 4. VEWS OF OTHERS. NA 5. RECOMMENDATON. Approve document for publc release. Sutablty s based solely on the document beng unclassfed not jeopardzng DoD nterests, and accurately protrayng offcal polcy. Karen S. Henry ProfessorffFCE 1 Tab AF lmt 1768, 19840901, VS PREVOUS EDTON WLL BE USED.
TECHNCAL NOTE Karen S. Henry, PhD, P E 1, TSgt Hed Hunter 2 and Benjamn Fonte 3 THE NFLUENCE OF PROCESSNG SOL WTH A COFFEE GRNDER ON SOL CLASSFCATON Abstract Use of a coffee grnder to break up clods of sols s one of the recommended practces of the Rapd Sols Analyss Kt (RSAK). Ths leads to the queston that t may produce fnes by cuttng/breakng larger partcles-potentally leadng to the ms-classfcaton of sol. n order to nvestgate ths hypothess, we performed a laboratory nvestgaton to determne f the use of a coffee grnder produces fnes from sand-sze partcles and whether the producton of more fnes can result n a dfferent uses classfcaton by some combnaton of addng sol fnes and changng the Atterberg Lmts of the fnes tested. Three sols were tested--ottawa Sand (SP), a poorly graded sand wth slt (SP-SM) and a sandy lean clay (CL). Processng the sols for 90 seconds n a coffee grnder produced from 15.9% to 18.5% fnes from the Ottawa Sand and establshed that the coffee grnder breaks down sand partcles nto fnes. For the coarse Elevator Sol, orgnally an SP-SM, the fnes producton changed the sol classfcaton to SM, whle the Lqud Lmt decreased from a measureable 21 to non-plastc, and Plastc Lmt decreased from 19 to 17. For the Harte Clay (CL), the sol classfcaton dd not change; but the Lqud Lmt and Plastc Lmt both sgnfcantly ncreased. Therefore, for both the producton of fnes and the resultng mpact on Atterberg Lmts, the nfluence of usng a coffee grnder to process sol cannot be quantfed wthout further, systematc study. 1. ntroducton The Rapd Sols Analyss Kt (RSAK) was recently developed by the US Army Corps of Engneers under the Jont Rapd Arfeld Constructon program (JRAC), to address the need for rapd, on-ste sol characterzaton (Berney and W ahls, 2008). The RSAK provdes sol 1 Professor, Department of Cvl and Envronmental Engneerng, Unted States Ar Force Academy, CO, USA 80840, Karen.henry@usafa.edu 2 Techncal Sargent, US Ar Force, Department of Cvl and Envronmental Engneerng, Unted States Ar Force Academy, CO, USA 80840, hed.hunter@usafa.edu 3 2" Lt, US Ar Force, Mnot Ar Force Base, Mnot, ND ljpage
classfcaton and constructon desgn parameters (e.g., Calforna Bearng Rato values) wthn a few hours, requres lttle experence on the part of the operators and can be used n place of the feld sol classfcaton technques such as those descrbed n the former Army Techncal Manual TM 5-410 (republshed as TM 3-34.64) n whch several tests requrng sklled techncans are used (rbbon test, jar test, etc..) (Berney and Wahls, 2008). t ncludes a small feld laboratory consstng of a mcrowave, electrc balance, seve shaker, seves, coffee grnder, plastc lmt tool, bowls, spatulas, and scoops. To classfy sols, a dry seve analyss s performed, as s a Plastc Lmt (PL) test. The gran sze dstrbuton nformaton and the PL (whch utlzes lnear regresson to predct the Plastcty ndex, or P) s then used to classfy the sol and predct Proctor compacton curves and soaked Calforna Bearng Rato (CBR) values as a functon of dry unt weght. The RSAK recommends the use of a small, hand-held coffee grnder to process fne-graned sols (.e., break down fnes that have clumped), descrbed as follows (Berney and Wahls, 2008): After mcrowave dryng of sols contanng a fne-graned fracton, "clods" of sol wll be present as a result of rapdly drvng out mosture from wet clumps of sol. Left alone, these clods would skew the gran sze curve to the coarse sde and create a false classfcaton of the sol, whch n turn would lead to mproper constructon gudance. Tradtonal tools to process sols are the mortar and pestle, whch can be used to grnd down hard clods nto much fner materal. The mortar and pestle also allows separaton of true aggregate from sol clods by observng ther response to grndng. Once a coarse separaton of the sol has occurred and any aggregate has been removed, there wll stll reman a large number of very fne clods that cannot be ground any fner wth the mortar and pestle. To reduce these fne clods to a powdered materal that can pass the No. 200 seve durng dry sevng, the RSAK ntroduces a coffee grnder that can pulverze these fne clods nto powder. As shown n the classfcaton secton of the report, the coffee grnder makes classfcaton of fat and lean clays much more feasble, whereas mortar and pestle refnement typcally results n a sandy-clay desgnaton. The suggested use of a coffee grnder leads to the queston that t may produce fnes by cuttng/breakng larger partcles-potentally leadng to the ms-classfcaton of sol. n early work, the US Army noted the mportance of usng the mortar and pestle properly to break down clods of fne-graned sol (TM 3-34.64, Chapter 5 regardng feld classfcaton of sol): 21Page
A rubber-faced or wooden pestle and a mxng bowl s recommended for pulverzng. Lumps may also be pulverzed by placng a porton of the sample on a frm, smooth surface and usng the foot to mash t. f an ron pestle s used for pulverzng, t wll break up the mneral grans and change the character of the sol; therefore, usng an ron pestle s dscouraged. As noted, the use of the hard pestle s purported to break sol partcles. We hypotheszed that use of a coffee grnder (wth steel blades) to process sol changes the gran sze dstrbuton due to cuttng and/or breakng of partcles. Furthermore, ths addton of fnes can change the Atterberg Lmts n at least some cases. n order to nvestgate ths hypothess, we performed a laboratory nvestgaton to determne f the use of a coffee grnder produces fnes from sand-sze partcles and whether the producton of more fnes can result n a dfferent uses classfcaton by some combnaton of addng sol fnes and changng the Atterberg Lmts of the fnes tested. The purpose of ths artcle s to document the results of the expermental nvestgaton. 2. Lterature revew/prevous work As noted above, early US Army gudance recommended usng only a rubber-faced or wooden pestle wth a mxng bowl to break up sol clods. Producng fnes by mproper sol processng has been a concern for a whle. An early laboratory manual for use n colleges and unverstes notes that a dsturbed sol sample s ready for testng when "all partcles smaller than the 40 mesh have been separated suffcently to pass the 40 mesh wthout fracturng the ndvdual grans;" and as regards to dry sol (Dawson, 1949). Kayabal (2011) tested the hypothess that because the Lqud Lmt and Plastc Lmt are drectly related to gran sze (.e., specfc surface area), Atterberg Lmt tests performed on the same sol would be hgher for the -#200 Seve fracton compared to the -#40 Seve fracton that s utlzed for the Unted Sol Classfcaton System (USCS) of sok He performed Atterberg Lmt tests on 60 samples of lacustrne clay (collected from dfferent locatons) -each par of tests comprsed one each of the -#200 Seve fracton and the -#40 Seve fracton of the same sol. He found that both the Lqud Lmt and the Plastc Lmt were on average 10%-20% hgher on the -#200 Seve fracton. Further, n 19 of the 60 cases, the USCS Classfcaton of the sol changed as a result of usng the dfferent fractons of sol. Hence, there s evdence that when the 3f Page
samplng and laboratory method used results n a fnes ncrease, ) the Atterberg Lmts ncrease and 2) ths sometmes changes the sol classfcaton. 3. Expermental desgn Our experments were desgned to test the hypothess that usng a coffee grnder to process dry sol produces sgnfcantly more sol fnes as compared to the standard use of mortar and pestle, as prescrbed by ASTM Standard D421-85, accordng to the Unted Sol Classfcaton System (USCS). Further, the addtons of fnes changes the Atterberg Lmts, and therefore the lkelhood of obtanng an accurate sol classfcaton s decreased. 3. Laboratory nvestgaton Ths study was conducted n two phases-) A determnaton of whether coffee grnder processng would break down sand-szed partcles nto fnes and 2) A study of the nfluence of the coffee grnder processng on the Atterberg Lmts of sol contanng fnes compared to standard mortar processng. 3.2 Phase : Does the coffee grnder break down larger sol partcles nto smaller partcles? n ths phase, sand-szed partcles were used. We began by processng 40-60 (.e., passes the Number 40, or 0.425 mm Seve, but s retaned on the Number 60, or 0.25 mm seve) Ottawa Sand n a coffee grnder and comparng t to unprocessed sand from the same sample to dscern whether the coffee grnder would produce fnes. Dry seve analyses (only) were performed for ths phase. The Ottawa Sand samples were processed by usng a sol spltter so that each control (no processng wth a coffee grnder) was matched by a sample that was processed n a coffee grnder. For the ground sample, we processed approxmately 600 grams of sand usng a new coffee grnder. The coffee grnder was only able to handle about 200 grams at a tme. Each 200 gram porton was ground for 90 seconds whle tltng the coffee grnder back and forth by hand (approxmately 30-45from vertcal n both drectons). The RSAK nstructons do not specfy a processng tme related to the use of the coffee grnder; however, we selected 90 seconds n order to treat each specmen n a smlar manner. After the three 200 g sub-samples were 41Page
processed through a new coffee grnder (90 sec each) they were recombned nto one sample. We used a 600 gram sample n order to optmze the accuracy of the gran-sze-dstrbuton determnatons. After the frst 90 second processng of one 200 g sub-sample of Ottawa Sand, the grnder blades showed notceable wear. Therefore we measured the grnder blade dmensons durng the study. Two locatons of the blade were measured: 1.) the wdth of the base of the blade closest to the center, and 2.) the length of the blade that forms an angle from horzontal (Fg. 1 ). The blades were measured before and after each tme sol was processed untl all 600 g were processed. Data analyss ncluded gran sze dstrbuton determnatons n a standard dry seve test for the Ottawa Sand processed n the grnder and the control sample of Ottawa Sand. 3.3 Phase 2 - Does coffee grnder processng change Atterberg Lmts and sol classfcaton? For ths phase of testng, two sols were utlzed --a sample collected from beneath the Southeastern elevator n Farchld Hall at the US Ar Force Academy, gven the name "Elevator Sol," whch classfes as SP-SM, Poorly graded sand wth slt, and a sol collected from Loveland, Colorado, gven the name "Harte Clay" (collected by Lt Sean Harte n 2009 from Loveland, Colorado), whch classfes as a CL, Sandy lean clay (contanng 42% sand and less than 0.5% gravel). Each set of sol samples (control and treated) was obtaned for testng va a sol spltter. For the control, a mortar and pestle were used to break up clumps of sol nto ndvdual partcles and the treated sol was processed wth the coffee grnder for 90 seconds as descrbed above. Seve analyss usng the wet preparaton method was used to test and classfy the sols after ther treatments wth mortar and pestle (control) or wth the coffee grnder. The Harte Clay had been stored ndoors for over two years, and was qute stff--t was n hardened chunks. Hence, a rubber mallet was used to carefully break apart large chunks of the clay pror to any treatment. Three hundred grams of the sol (total) were used for the coffee S Page
grnder processng-and we used fve 90 second cycles of processng approxmately 60 g of sol. Measurements of the blades of the coffee grnder were taken before any sol was processed and after the fve cycles of processng sols (as was done for the Ottawa Sand). Expermental Results 4.1 Wear on Coffee Grnder Blades due to sol processng The base wdth of new blades measured 10.1 mm, and the length of the new blade tps was 6.4 mm. The wear on the blade tp wdth (Fgure 1, left sde) was less than 1 % of the orgnal wdth for all measurements made. The length of the blades; however, had notable wear, dependng on the sol processed, as shown n (Table 1 ). The Elevator Sol wore down the blades much more quckly than the other sols tested. Fgures 2 and 3 show the pctures that were taken after one processng cycle and three cycles, respectvely, of Elevator Sol. The Harte Clay had very lttle effect on the wear of the blades. A total of fve cycles of the Harte Clay were run through the grnder and the grnder had less wear on t than wth only one cycle of Elevator Sol. The Ottawa Sand eroded the blades sgnfcantly, but not as much at the Elevator Sol. Table 1- Measurements of coffee grnder blades after processng sol Sol Number of90 % decrease n tp second cycles length Ottawa Sand 4 21 Elevator Sol 1 10 2 28 3 100 (no tp left) Harte Clay 5 2 G Page
Fgure 1: Top vew of new RSAK coffee grnder, showng where blade dmensons were measured before and after processng sols. Fgure 2: The blades after one 90 second cycle of Elevator Sol Fgure 3: The blades after three 90 second cycles of Elevator Sol 71Page
4.2 Ottawa Sand Dry Seve Analyss-Does coffee grnder processng produce fnes? The Ottawa Sand (control) was 99.8% sand wth the remanng porton beng fnes (Fg 4). t was very well-sorted, the Coeffcent of Unformty (Cu) was 1.95 and the Coeffcent of Curvature, (C,) was 0.95. The sol classfes as SP, Poorly Graded Sand, accordng to the Unfed Sol Classfcaton System (USCS). The frst sample of Ottawa Sand that was processed wth the coffee grnder resulted n 83.9% Sand wth the remanng porton beng fnes (Fg. 4). The Coeffcents of Unformty and Curvature were not determned, snce D 10 was smaller than the openng sze of the No. 200 Seve (0.075 mm). Hence, the sol would ether classfy as SC or SM. Presumably the fnes produced are non-plastc, and the sol would classfy as SM. Three more samples of Ottawa Sand were processed wth the coffee grnder, and the percentages of fnes n each sample are shown n Fgure 5, and ndcate that the percentage of fnes produced are relatvely consstent, wth values rangng from 15.9% to 18.5%. g 70 60 80 -+-Ottawa Sand Ottawa Sand After Grnder Processng 50 f------+---+--+--+-t--++--_,_+-lf----+-+-+-f----+--- j ; 40 f-----;---+--+---+-++-c-t----ott-1-r--+-+--+--+-h-----c " 30 f-----r--+----+-+-c--c-;-+--t----j, f---+-+-+--+--+--- 20 10 0.01 0.10 1.00 Seve Sze (mm) Fgure 4: Gran sze dstrbuton of Ottawa Sand and Ottawa Sand after beng processed n a coffee grnder for 90 seconds S Page
ll " 15 =... 10 g..... " ; 5 ll 0.. 0 o.z Control After 1 After 2 After 3 After 4 Fgure 5: Percentage Ottawa sand passng the No. 200 Seve for untreated sol (control) and after processng wth the RSAK coffee grnder. After 1, means that ths was frst tme that the coffee grnder was used, After 2 means that the coffee grnder had already processed one batch of sol (so that there was some wear on the blades); smlarly for After 3 and After 4. 4.3 nfluence of coffee grnder processng on gran sze dstrbuton, Atterberg Lmts and sol classfcaton of the Elevator Sol and Harte Clay The gran sze dstrbutons for the control and treated samples of both the Elevator Sol and the Harte Clay are shown n Fgures 6 and 7, respectvely; and Fgure 8 s a bar graph showng the percentage of fnes for the control and processed specmens. The Elevator Sol contaned 4% gravel, 87% sand and 9% fnes. For coffee grnder processng, the same sol was found to contan no gravel, 72% sand and 29% fnes-thus, the coarse graned sol content decreased by 19%, whle the fnes content ncreased by the same amount due to processng. For the Harte Clay, the fnes ncreased from 57% to 64%. n each case, the coffee grnder processng resulted n the producton of sgnfcant fnes compared to the control specmen. For the Elevator Sol, the processng reduced the coarse materal content to 77% of that of the orgnal value and for the Harte Clay, the coarse partcle content was reduced to 84% of the orgnal value. The fnes produced were enough to change the sol classfcaton of the Elevator Sol from a dualclassfcaton to that of Slty Sand (SM). 9Page
100 90 80 70 #... c 60....... ".. 0. so c 40 " u a:. 30 20 f-- -.-Elevator... Elevator - Grnder v j...---- y 10 " 0 0.01 0.10 1.00 10.00./v / / : : / / }, X \ :, V / y /,/ \. / / J.. / Seve Sze (mm) 1 Fgure 6: Gran Sze Dstrbutons for Elevator Sol and Elevator Sol processed wth a coffee grnder. 100 90 80 70.. c 60 ;;; m 0. 50.. " ) c " 0. 40 30 20 10... /( ){" v.v: -.-Harte \ Harte - Grnder -, v l*f,. :, :., ; 0 0.01 0.10 1.00 10.00 Seve Sze {mm) Fgure 7: Gran Sze Dstrbutons for Harte Clay and Harte Clay processed wth a coffee grnder.. 10 Page
70 60 50 " c -= 40... " "" 30 c " " u 1. 20 10 0 Harte Clay Control Harte Clay Elevator Sol Elevator Sol Grnder Control Grnder Fgure 8: Percentage of the fnes n the control and treated (after processng wth the RSAK coffee grnder) samples of Harte Clay and Elevator Sol. Table 2 lsts the Atterberg Lmts and the USCS classfcatons for matchng control and coffeegrnder-processed specmens. For the Elevator Sol, orgnally classfed at SP-SM, the Lqud Lmt could not be determned on the processed sol usng the Casagrande procedure because the fnes were non-plastc, and the Plastc Lmt of the materal passng the Number 40 seve decreased slghtly. The fnes classfy as ML for both the control and treated specmens. Table 2: nfluence of coffee grnder processng on the Atterberg Lmts of Elevator Sol and Harte Clay Sol Lqud Lmt (%) Plastc Lmt (%) Plastcty ndex uses Control Grnder Control Grnder Control Grnder Control Grnder Ottawa Not applcable to control, dd not perform on grnder processed Sand sol. Assumed that the fnes would be non-plastc. SP SM Elevator Non- Non- Sol 21 plastc 19 17 2 plastc SP-SM SM Harte Clay 37 41 14 25 22 26 CL CL For the Harte Clay, both the Lqud Lmt and the Plastc Lmt ncreased due to processng the sol wth the coffee grnder. The Lqud Lmt ncreased from 36 to 41, and the Plastc Lmt 111Page
ncreased from 14 to 25. (Thus, the Plastcty ndex decreased.) Whle the USCS classfcaton of the sol dd not change n ths case, t suggests that for fne-graned sols that fall close to the A-lne of the Plastcty Chart, the sol classfcaton could change. 4. Dscusson The results of ths research program support our hypothess that the use of a coffee grnder to process sol changes the gran sze dstrbuton, and that the addton of fnes can change the Atterberg Lmts n at least some cases; both of whch can result n a dfferent sol classfcaton. For the coarse Elevator Sol, orgnally an SP-SM, the fnes producton changed the sol classfcaton to SM, whle the Lqud Lmt decreased from a measureable 21 to non-plastc, and Plastc Lmt decreased from 19 to 17. For the Harte Clay (CL), the sol classfcaton dd not change; but the Lqud Lmt and Plastc Lmt both sgnfcantly ncreased. Therefore, for both the producton of fnes and the resultng mpact on Atterberg Lmts, the nfluence of usng a coffee grnder to process sol cannot be quantfed wthout further, systematc study. A lkely explanaton for the Atterberg Lmts to decrease for the Elevator Sol, s the addton of fnes due to breakng of sand partcles, whch are non-plastc. The results algn wth those of Kayabal (2011). n the case of the Harte Clay, the coffee grnder may have broken not just sand partcles, but clay partcles, thus ncreasng surface area avalable to react wth water. nterestngly, the RSAK sometmes under-predcts the amount of fnes n sols (Berney, et al, 2008). "...t s noted that, n all cohesve sol cases, the RSAK predcts a lower fne content as a drect result of not mostenng the sol as per ASTM C117. ASTM Cl 17 states that only when the accuracy of materal passng the No. 200 seve s requred wash sevng should be used. The use of the coffee grnder s an attempt to mtgate ths dscrepancy, provdng a means to generate suffcent fnes quantty to dfferentate fne-graned sols from coarse-graned sols. Hence, the lnk between the use of the coffee grnder, whch ncreases fnes, and the small, dry seve stack, whch results n a under-predcton of fnes mght be an area of further study. 121Page
We emphasze that ths lmted research effort was ntended to test the hypothess descrbed above, and t s not a crtcal nvestgaton n regards to the valdty of the RSAK. For example, although we processed each sol for 90 seconds n the coffee grnder, the RSAK does not specfy an amount of tme (Wahls 2008): To begn sol processng, the dred sample from the prevous step s pulverzed wth a mortar and pestle and coffee grnder wth care taken not to break up soft aggregate, but only sol clods that have formed durng mcrowave dryng. t s very lkely that n feld use to process sols, the coffee grnder would be used for a shorter amount of tme than 90 seconds, and would produce fewer fnes than those produced n our study. n addton, the RSAK recommends the use of the coffee grnder to process clay sol, whereas we used t to process sands n order to test our hypothess. Fnally, the wear on the blades of the coffee grnders due to sol processng ndcates that for feld use, the coffee grnder has a fnte lfe span. t s logcal to assume that the blades wll contnue to wear down untl the coffee grnder can no longer break down the sol partcles. 5. Conclusons Based on the experments conducted, we conclude that: 1. Usng the processor wth the Ottawa Sand establshes that the blades break down sandszed partcles nto fnes. The 90-second processng tme resulted n the producton of 15.9% to 18.5% of fnes from the sand partcles. 2. The amount of fnes produced by 90 seconds of processng n a coffee grnder was enough to change the Unted Sol Classfcaton System classfcaton of and poorlygraded slt sand (SP-SM) to slty sand (SM). 3. The amount of fnes produced by processng wth a coffee grnder resulted n a reducton of coarse partcles to 77% of the orgnal value for the orgnal SP-SM tested, and to 84% of the orgnal value for the CL tested. 4. The Atterberg Lmts determned for both treated sols-the Elevator Sol and the Harte Clay were dfferent than those determned for the control specmens of each sol. They Bl Page
decreased for the Elevator Sol and ncreased for the Harte Clay, suggestng that further study would be requred to be able to adjust for the use of the grnder. 5. The coffee grnder blades wear down when processng sol. The degree of wear wth each use vares dependng on specfc sol type, suggestng a lmted lfespan for feld use. Acknowlegements The authors wsh to acknowledge the contrbutons of Leutenants Kevn Doolttle and Erc Glpatrck, who helped establsh the expermental procedures descrbed n ths paper. References ASTM D421-85. Standard Practce for Dry Preparaton of Sol Samples for Partcle-Sze Analyss and Determnaton of Sol Constants Berney, E., and Wahl, R. (2008). A Rapd Sols Analyss Kt, Unted States Army Corps of Engneers Engneer Research and Development Center, ERDC/GSL TR 08-3, 119 p.: http://www.dtc.ml/cg-bn/gettrdoc? AD=ADA484 l 59 Kayabal, K. (2011) Determnaton of Consstency Lmts: A Comparson between -#40 and - #200 Materals, Electronc Journal of Geotechncal Engneerng, Vol. 16, Bundle T, pp. 1547-1561: www.ejge.com/2011/pprl 1.172/Pprl 1.172alr.pdf U.S. Army Techncal Manual TM 3-34.64 (FM 5-410/23 Dec 1992)/MCRP 3-17.70. Mltary Sols Engneerng, Headquarters, Department of the Army, Washngton, D.C., 456 p. 14 Page