XI B RARY OF THL UN IVLRSITY OF ILLINOIS 666 Ho. 16-23 CENTRAL CIRCULATION AND BOOKSTACKS The person borrowing this material is re- sponsible for its renewal or return before the Latest Date stamped below. You may be charged a minimum fee of $75.00 for each non-returned or lost item. Theft, mutilation, or defacement of library material* can be causes for student disciplinary action. All materials owned by the University of Illinois Library are the property of the State of Illinois and are protected by Article 16B of lllinoit Criminal law and Procedure. TO RENEW, CALL (217) 333-8400. University of Illinois Library at Urbana-Champaign $f P 1 « ?0(fl MAY 2 9 2001 When renewing by phone, write new due date below previous due date. L162 UNIVERSITY OF ILLINOIS BULLETIN I ssr HI) WEEKLY Vol. XI. MAY 11. I'M ( No [Entered as second-class matter December II, I'M.', at the post offi L'rtiana, Illinois, under the A ist 24. 1912.] BULLETIN No. 19 DEPARTMENT OF CERAMICS R. T. STULL, Acting Director INVESTIGATION OX IRON ORK CEMENTS BY ARTHUR JE. WILLIAMS PUBLISHED BY THE UNIVERSITY OF ILLINOIS, URBANA 19 13 I 9 V4 Authorised Reprint from the Copyrighted i Volume \ 111. 1012 NATIONAL ASSOCIATION OF CEMENT USERS. Pill! MU I I Ml K. I'lWI [RON ORE CEMENTS By Aktih h E. \\ ii.i.i wis.t Iron ore cement is a producl intended to be used in sea water work. This material is now manufactured in Europe under tin- name of Erz cement. According to Mr. William Michaelis, Jr.,J '•='- the process of manufacture is similar to that of Portland cement except that limestone and iron ore axe used in place of limestone and clay. United States Consul Thackara§ gives a description of its manufacture as follows: Chalk, flintstone, and finely ground ferric oxide are used. The hint and iron are ground together, then mixed with the chalk and water and BCreened through a fine sieve. The screened producl is clinkered in a rotary kiln and then ground. An average composition of iron ore cement, given by Michaelis is: CaO 63.5 per cenl Al : <> 3 l 5 per eenl Si0 2 20.5 " Mn<> 1 5 FejOj 11.0 " Alkali 1.0 The effect of sea water is undoubtedly two-fold. In the firsi place chemical reaction may take place between certain con- stituents of the cement and the salts in sea water, and. on the other hand, the mechanical action of the waves carrying large amounts of sand, freezing, thawing, and the varying pressure of the water due to tide help to injure the cement BUbmerged in sea water. This work, however, will be confined to the chemical action of sea water, for the mechanical action is of minor import- ance unless the cement is weakened by chemical changl The reactions which take place between Portland cement and sea water are said to be of three distinct kinds. First, t In- action of MgCla and MgSt ) 4 in sea water on the calcium hydrate formed during the hardening process <>f the cement, forming Mg(OH) a , CaCl 2 , and CaS0 4 . Second, the action of gypsum, * Under tin- direction of Mr R T Stull. t Orbana, 111. A Thesis for the Bftcheloi ol - Illinois in 1910. X Eng. Sms, Vol :■>.. pp. 545 646 \ United Statin Consular Reports, June, 1008. (1) 2 Williams on Iron Ore Cement. CaS0 4 formed above, upon the calcium aluminates forming calcium sulpho aluminate. Third, the crystallization of the gypsum and calcium sulpho aluminate giving an increase in volume, thus causing the disintegration of the mortar. That free lime is present in set Portland cements is well known. Lamine* found 32 per cent of CaO in cement sub- merged in the Black Sea 15 years. Every analysis of a cement exposed to sea water shows a high percentage of MgO. Yicatf in 1840 showed this fact clearly, a cement, which was submerged in sea water for 6 months, was analyzed. A sample, taken from the surface exposed to the sea, showed 10.4 per cent MgO and 19.3 per cent CaO while the interior, which was not impaired, showed 1.87 per cent MgO and 31.33 per cent CaO. A. MeyerJ states that cement loses strength in sea water. The MgS0 4 acting with the silicate of lime forms Mg(OH) 2 and calcium sulphate. The CaSO^ reacts with the calcium aluminates (A1 2 3 , x CaO) of the cement, forming Al(OH) 3 + 3 Mg(OH) 2 + CaS0 4 + CaCl 2 . Charles J. Potter§ says that MgS0 4 is the most active con- stituent in sea water on cement. He found that MgCl 2 softens cement but causes no expansion. Potter says that it is now definitely believed that magnesium salts act on the feebly com- bined lime and alumina compounds which on taking up water of crystallization cause bursting of the concrete. He mixed calcined red brick clay with Portland cement clinker in propor- tions of 6 to 10. From this mixture briquettes were made and placed, together with Portland cement briquettes, in fresh water, sea water, and sea water to which 10 per cent MgS0 4 was added. Both of these cements gained strength in fresh water. In salt water, the Portland cement briquettes began to fail after 5 weeks and were disintegrated after 5 years. These cements showed blistering after one year, which was followed by expansion and bursting. The red cement improved continually but took 8 weeks to obtain the maximum strength that the Portland cement had obtained in 5 weeks. In the 10 per cent solution of MgS0 4 , the Portland cement tested 500 lb. in a month and then went * Le Ciment, 1901, pp. 111-691-81. t Iron Ore Cement — The P. C. Co. of Hemmoor, Hamburg, Germany. t Chemisches Central Blatt, Vol. 73, p. 1368. § Jour. Soc. Chem. Ind., Vol. 28. Williams on Ikon Orb CEMENT. back to zero in 1 year. The red cement began at 250 lb. and increased continually to 1015 lb. in 8 years. Mr. Potter says thai the chemical combination <>f CaO, SiOj, and Al<» and water is feeble and that probably accounts for the ability of magnesium in sea water to be bo active. The experiments of Dr. Miehaelis* and l.e Chatelierf lead them to the conclusion thai Portland cemenl Buffers in solutions containing sulphuric acid salts, which applies to sea water. A douhle salt is formed composed of gypsum and calcium aluminate. This sulpho-alunhnate. AUh. CaO + 3CaSC>4, is said to crystal- lize with 30 molecules of water, which process musl be accom- panied by considerable expansion. Le ('hatcher Bays that "the main cause it' not the sole cause, of the injuries which cements suffer under the action of sea water is the formation of calcium sulpho-alunhnate." RebuffatJ says on the contrary that sulpho-aluminates cannot exisl in cements in sea water hut agrees with Miehaelis and Le (hatcher that calcium ahmiinates are the part- of cement most easily acted upon by salt- in sea water. It has been shown that calcium ferrates are formed similarly to the calcium aluminates and that alumina could he replaced by ferric oxide in Portland cement. Dr. Miehaelis puts this knowledge into use with the idea of overcoming the disintegra- tion in sea water. The result of this application is the Iron Ore cement of today. Dr. Miehaelis and the Royal Experiment Station of Charlot- tenburg have tested these cements in comparison with Portland cements in a very thorough manner. Mr. William Michaelis§ -ay- in a paper read in the United States that tests of Erz cement and Portland cement were made with both neat and '■'> to I mix- tures which were placed in fresh water, sea water, and water containing five time- more -alt that sea water. In -ea water. the Krz cement developed a much greater strength than the Portland. In the strong -alt water, the strength of the Portland cemenl decreased rapidly while the Irx cemenl showed a steady gain. Briquettes were made of Iron Ore and Portland cement * Ton /• to I ■•. L896, p. 838. i I . Ciment, 1901, p 31 32 : T •/. 1901, i> -'72. § £Viy. Newt, Vol. 58, pp 645 646 4 Williams on Iron Ore Cement. which were placed in a salt solution of five times the normal strength of sea water under pressure of 15 atmospheres for a few days. This condition destroyed the Portland cement bri- quettes entirely, while the Iron Ore cement increased in strength. The Royal Experiment Station conducted similar tests to the above but much more elaborate. Two Iron Ore and three Portland cements were made into prisms, using a 3 to 1 mixture of standard sand and cement. These prisms were placed in sea water and water containing five times the percentage of salts in ordinary sea water. In addition to this, these three solutions were allowed to act upon test pieces made of cement mixed with varied amounts of gypsum. All the Portland cement mortars disintegrated in the three- and five-fold salt solutions; all the Iron Ore cement mortars remained intact and sound. United States Consul A. W. Thackara* investigated this cement for use on the Panama Canal. The result of his investi- gations was the adoption of this cement for concrete work exposed to sea water. Another point in favor of this cement is the property of slower setting. The cement is weaker than Portland for the first week, but then gradually gains strength and exceeds that of Portland. Publications of previous experiments do not show definitely the best composition for cements giving the greatest protection against sea water. With this idea in view, the following investi- gations were undertaken: The outline of procedure in these experiments is as follows: Newberry's cement formula, x (3CaO, Si0 2 ) + y(2CaO, A1 2 3 ), was used as a basis. Assuming, according to Xewberry, that Fe 2 3 could replace A1 ? 3 and form 2CaO, Fe 2 3 , a triaxial dia- gram was plotted (Fig. 1), the three members stationed at the three corners being 3CaO, Si0 2 , 2CaO, A1 2 3 and 2CaO, Fe 2 3 . By blending these three members, cements could be obtained containing various amounts of the calcium aluminate and the calcium ferrate. The batch weights of these three members were calculated and about 15 kg. of each were weighed up, using practically chemically pure materials. Whiting, flint, aluminium hydrate, and red oxide of iron were the only ingredients. These batches * United States Consular and Trade Reports, June, 190S. Williams on Leon < >re < Iemen i . were ground in a ball mill, then passed through a 200-mesh Bieve; thus getting thorough mixing and a finely ground batch. The formula 1 for the cements made are given in Table I. The following cements, No. 19, 20, 21, 22, 23, 24, 25, 36, 37, 38, 39, 40, 42. is. 49, 50, 51, 52, 53, 54, 58, 59, 60, 61 62, and 05 on triaxial diagram were then weighed up, blunged thor- oughly, and partially dried by pouring the slip into plaster molds. / - Jo A. Aio 2SJ A «fe h ft / ^ / 4S\l0 <%>/ / "f\ */\ • • "/\ */\ "/ \ /\ /\ o o/l 2) 22 J3. ■± :/ 6$. di/ as/ oa\ FIG. 1. — TRIAXIAL DIM. HAM. The cements were then rolled into small balls aboul the size of a marble, dried, dehydrated in a down draft kiln to about Sl "> C. and placed in fruit jars ready for burning. These cements were burnl in a magnesite tesl kiln, designed by Mr. Stull of the Ceramic Department, especially for burning experimental cements. The construction of this kiln ifl Bhown in Fig. 2. The success of this kiln is a noteworthy fact ae 6 Williams on Iron Ore Cement. kilns suitable for this purpose, heretofore, have not been very satisfactory owing to lack of control, unevenness of temperature in the clinkering chamber. Kerosene oil was used for fuel with an air pressure of about 50 lb. The temperature at the time the clinker was drawn from the kiln was determined first by means of a "Wanner pyrometer. This was given up, however, as the rapid rate of burning required a higher temperature than the true temperature of clinker forma- tion. Table I. — Formulae of Cements Made. No. Formula. Molecular Ratio SiOiiAlO+FeiOa 19 20 21 22 .l(3CaO,Si0 2 ) +.2(2Ca0,Al20 3 ) +.7(2Ca0,Fe 2 3 ) 0.11 .l(3Ca0,Si0 2 ) + 1 (2CaO, AI1O3) + .8(2CaO,Fe 2 3 ) 0.11 l(3CaO SiOj) + O12CaO.Ff.O3) 0.11 .2(3CaO.Si02) +.8(2CaO,Fe 2 3 ) 0.25 23 .2(3CaO,Si02)+.l(2Ca().AM) 3 )+.7(2CaO.Fc203) 0.25 24 .2(3CaO.SiO-)+.2<2CaOAb0 3 > 4-.6(2CaO,FeiOa) 0.25 25 .2(3CaO.Si0 2 ) +.3(2Ca< '.Alt >„> - 5 2Ca< >.F«- ■■< >.-,> 0.25 36 3(3CaO.SiOi) +.2(2CaO '■ >(2CaO.Fet< 0.43 37 .3(3CaO.Si02) 4-.l(2Ca< ».AK)a) + .6(2Ca0.Fe 2 3 ) 0.43 38 .3(3CaO.SiO-) +.7(2CaO,Fe»0 3 ) 0.43 39 .4(3Ca0.Si0 2 ) +.6(2CaO,FejOa) (i hi', 40 .4(3Ca0,Si0 2 )+.l(2Ca< i.AljO - 5 2Ca< »,Fe»Oa) 0.66 42 .4(3CaO,SiOt) +.3(2Ca< >.A1 ■< 1 I +.3(2Ca< I.FesOa).. . 0.66 48 .5(3Ca< l,SiO h.3(2CaO,AhO +.2(2CaO,l - I I 1.00 49 .5(3CaO.Si0 2 t +2 2< aO \ . ". +.3(2CaO.Fe»Oa) 1.00 50 .5(3CaO,SiOa)+.l(2CaO,AljO - 1 !CaO,FejOa) . . . 1.00 51 .5(3CaO,SiOj) 4-.5(2Ca< >.¥<■< la) 1.00 52 .6(3CaO,SiO») 4-.4(2Ca< >.F.- < > • 1.50 53 54 58 59 60 61 62 65 .6(3CaO,SiOi) 4-.l(2Ca< i.Al-.Os) 4-.3(2CaO,FetO») 1.50 .6(3CaO,SiOt) +.2(2CaO,Al 2 0») +.2(2CaO.] 1.50 .7(5CaO.SiOj)+.2f2CaO.A! Oa + 1 2CaO.Fe*0 1 2.33 .7(3Ca0.Si02) +.l(2Ca( >.A1 ( ) 1 +.2(2Ca< >.Fe203) 2.33 .7(3Ca0,Si0s) +.3(2CaO.F. 1 t } ) 2.33 O.Si02) +.2(2CaO.Fej< I - 4.00 .8(3CaO.Si0 2 ) -r .1 2CaO,Al20 1 + 1 JCaO.FeaOa) 4.00 .9(3CaO,Si02)+.l(2CaO,Fe20 3 ) 9.00 Almost all of these cements were fused till the surface was glassy in appearance before the cement seemed well clinkered and crystals appeared. Cements No. 54, 58, 62, and 65 appeared like a Portland clinker, except darker in color and were not fused or slag-like in appearance. The clinker was first reduced in a jaw crusher and then ground in a disc mill; a screen test showed 24.2 per cent on 150 mesh screen; 12.3 per cent on 200 mesh screen; and the remainder, 63.5 per cent passed 200 mesh. These cements show that the}' are approximately of the same degree of fineness as the average Portlands. After the samples were ground, pats were made from Williams on [bon Obi Cement. '" ^- J : - ! ! H- 1 1 £*js " ";tfl f: ? i i 8 Williams on Iron Ore Cement. them in the usual manner to determine the properties of the cement. The amount of water used for mortar was determined by the Boulonge method (Waterbury's Cement Manual, p. 44). The initial and final sets were determined with Gilmore needles. Four pats were made of each cement with the idea of using one for the time of setting tests and placing the other three imme- diately in the moist closet, two of which were to be used for the boiling test after 24 hours, the third to be allowed to stand in Table II. — Results of Tests on Cements. No. 19 20 21 22 23 24 25 36 37 38 39 40 42 48 49 50 51 52 53 54 58 59 60 61 62 65 Time of Initial Set, hours. IX 1 2X 1% 1 X 1M 1H l l 2 IX X IX ix 3 2 1 \X 1 1 X IX l IX IX Time of Final Set, hours. Water Used, per cent. 3 5 5X 4 5H ii 2X 5 8 3X 2% 7 3 io 9 4 4X 3X 4X 5 6 21.0 20.0 21.0 20.0 21.0 22.0 21.5 20.0 20.0 20.0 21.0 20.0 21.5 22.0 21.0 22.0 21.0 20.0 21.0 23.5 22.0 21.0 21.0 22.0 22.0 21.0 Remarks at Time of Final Set. Cracked in M hour ( >. K. Strong No cracks Small cracks Cracked Cracked Cracked Cracked Cracked O. K. Cracked Cracked O. K. O. K. Cracked Cracked O.K. Soft Cracked Cracked No crack? ( >. K. Soft and crumbly Warped Did not harden Cracked Conditions after 48 Hours in Moist Closet. Cracked Warped and cracked N<> cracks No cracks No i racks No cracks No cracks. O. K. No cracks No cracks Cracked Warped No cracks No cracks Cracked No cracks. Soft Soft No cracks. No cracks. Cracked Warped Warped and cracked O. K. (). K. Warped Soft Soft o. K O. K. water for 28 days. All of these cements went to pieces in cold water or in the boiling test. The results are given in Table II. From these cements, one only, i. e., No. 62, remained sound when placed in water. This cement also stood the boiling test {\ hr.), the others going to pieces. The molecular ratio of Si0 2 to AI2O3 for this cement is four and since the molecular ratio for good cements is between 5.1 and 6.8 and since none of these cements lie between these limits, it was decided to construct a new group. Cement No. 62 approached these ratios nearer than any other. Williams on [ron « )ke Cement. 9 A new batch \\;is calculated after Bleininger'a formula (2.8CaO,Si0 2 ) H (2CaO, AJA) having different amounts ol FegOa and Al,o, and also the ratio of SiOj to AU>, | I varied from jusl above to jusl below the limit-. The using ol chemically pure raw materials in place of Blag and limestone gives less efficient mixtures of lime andSiOj. It was, therefore, thought that sufficient lime would be obtained by the use <>t Bleininger'a formula. For formulae Bee Table III. Table III. — Formulae fob Cements Mad No. .1, .1 \ Bi 1'.. n I.. C, Ci Formulae :, i .' BCaO.Sil I » FeO I 5 8 -' B< ':.' ».Si< »«) • -'< !a( »,Fi 6 l(2.8Ca< '.Si< It) 4-(2Ca< l.FesO 7 n 2 8Ca< ».Si< I.' - C2Ca< 1,1 ■ 5.25(2.8CaO,SiO L75 21 O.AW3 B25(2I O.FejO 6.00(2.8CaO,SiO • l75(2CaO,AltOi) • B25(2Ca0.F< 6 in 2 8Ca< >,Si< > - 200 2< !a< >. \H ' BOO -'< la< >,FeiO 7 22 .' BCa< >.Si( h) -hl75(2Ca< »,A1»0 • 825(2Ca< ».FejO .-, ii 2.8< la< I.SiO 1(21 ' I \i iO I ■ 640 -'< !a< ».F( 5 v., 2 BCaO.SiOil I WO 2CaO, \1J 2B O.F< 6 in 2 8Ca< >.Si< i • MM 2CaO,Al»Oi) • 2< O.FejO 7.00 -' SCaO.SiOi • 100 2Ca< '. \ I I 800 -'< ' l.FetO I'l.li. is I USE COMPOSn Mis Molecular Xo. ( .., , AltOi FejOi Ratio RjO 3iO A, 68.0 0.0 Hi. 22 i :. i At 66.7 0.0 in 1 22 9 5 B Az 67.2 II II 9 ii 6 i At 67 5 tl s 9 23 6 7 ii H, 66 7 1.3 g i Bt 67 I 11 v 1 •-':< l 6.00 Bt 67 5 l .3 7 -v 23 i 6 in Hi 68 l ii M - 2 Ci 67 i 2 ■". 7 2 .... i, 6 ii C: 68 2 7 6 ii Ci 68 2 2 .'. :, 8 .. in - 68 "• 2 . 3 :. 1 7 MM These Cements were prepared ill the same manner except that the temperature of clinkering was determined as uear as possible by the method used. The kiln was allowed to cool to about 1000 deg. C. before a batch of cement was pu1 in and tem- perature was then gradually raised till clinker wa- formed, the temperature was then read with a Wanner pyrometer. The clinkers obtained appeared exceptionally good, being dull black in color and glistening brightly in the Bun. These 10 Williams on Iron Ore Cement. clinkers were pulverized the same as has been previously de- scribed, then tested. The results of these tests, Table IV, show that good cements can be obtained with a large amount of alumina using the same ratio of Si0 2 to R 2 3 as Portland cements require. One very noticeable fact, however, is that when no Al>0 3 is present as in series A, A 2 , A s , and A± these cements all show expansion, thus giving evidence of free lime. Although A\ stood the boiling test, the cubes made from this cement bulged out from the mold considerably. The question arises at this point, is it always necessary for A1 2 3 to be present or can a good cement be made without it? Table IV. — Results of Test. Temperature Time to Clinker, hours. when Appearance Initial Set. Final Set, H2O, No. Clinkered, deg. C. of C'iinker. hours. hours. per cent. .-li 1300 :1 4 24 62 24.8 As 1320 Yi All 22 56 24.0 As 1320 m clinkered 26 56 23.2 .44 1330 l A good, 28 60 26.0 Bx 1390 Vl colored black 4M 40 26.3 B2 1320 Wi and 4H 44 24.4 B 3 1350 % glistening 11 36 28.0 Bt 1400 IH with 5 48 IT, O Ci 1320 14 crystals 5 30 24 . 4 d 1320 ; , in a 12 40 24.0 d 1330 1 ; , bright 12 48 28.0 Ca 1380 M light 17 40 27.2 This ought to be possible by reducing the lime content, as A x was the best of series A and also had the smallest amount of lime silicate. The slowness of setting is another factor which must be considered. It will be seen by Table IV that all of the cements required a long time to harden. This must be carried on in a moist atmosphere also or the cement will dry out before it has completely hydrated and set. The above factors will perhaps limit the use of this cement to work under water which may be allowed to set a considerable time. All the cements of series B stood a 6-hr. boiling test with- out showing any signs of expansion. In series C all but C\ stood the boiling test, d warped a little and came loose from the glass Williams on Ikon < >ke < i \n \ i. 11 plate although the cemenl has a comparatively l«»w lime content and its formula lies between other good cements. The attempt was next made to give these cements a com- parative test with Portland cement i" -how their relative resist- ance to sea water. The method used was Bimilar to that <>f Dr. Michaelis. One-inch cubes were made of each series oi cements together &^ FIG. 3. — sTKAM CYLINDER. with a 8e1 of cubes of a standard commercial Portland cement, which had stood all the commercial tests. These wen- allowed to stand <»<) hr. in the moist chamber and then placed in water, remaining in water for 27 days. The cubes made from series .1 together with a set of 5 Portland cement cubes were placed in a steam cylinder. Fig. :•;. containing an artificial Bea water Bolution of ten times normal strength. The quantity of salt is shown in Table V. The cement- were then put under steam pressure 12 Williams ox Irox Ore Cement. of 125 lb. or 8| atmospheres, the temperature being between 150 and 200 cleg. C. This was continued for 3 days. On opening the cylinder, the salt solution was found to be very dilute due to condensation of steam and no visible action on the cements had occurred. The salt solution and cubes were then put into a large wide-mouthed bottle, provided with a stopper and small vent hole. The bottle was then placed inside the pressure cylinder and steam admitted, allowing little or no condensation. After being sure that the bottle was not broken by the first change in temperature, the pressure was kept on for 3 days longer. Upon opening the cylinder, the cubes were found bone dry and covered with salt and the bottle cracked. This was due, no doubt, to the rapid reduction of the pressure, allowing the water T \ble V. — Analysis OF Ska Water.* Salt. Per cent of Suit. Ten times per cent of Salt. Total for 12 liters of Water. XaCl MgCli MgSOj CaS0 4 K-SO4 77 7" 10.87 4 . 73 3.60 2.46 0.217 0.345 108.7 47.3 36.0 342.10 478.28 208.12 158. 10 10.80 MgBr CaHC0 3 0.93 1.62 37.3 parts per thousand parts water. 100 parts =2700 parts water. 12000 =4.4 factor times per cent of salt = quantity per 12 liters of water. 2700 to vaporize rapidly, which was at a temperature above its boiling point. The results of this test were contrary to what was expected as the Portland cements were untouched and all of the iron cements were cracked and swollen. This cracking and swelling is caused, no doubt, by an excess of free lime, as these cements showed an expansion in the boiling test and there was a deposit of hydrated lime in the bottom of the cylinder which seemed to have been leached out of the cubes. Xo crushing strength test of Series A was made as they were all destroyed already. Series B was then placed in the cylinder, with a set of Port- * University Geological Survey of Kansis, Vol. 7, p. 27. Williams on Iron I mm: I !emen i . 13 land cement cubes. A vessel made of l-in. pipe was used in place of the glass bottle to overcome cracking due to suddeu change in temperature. This series was kept under pressure for 6 days, and when removed from the cylinder neither the Portland or Iron Ore cements appeared harmed excepl cemenl I!., which wtnt to pieces. The reason for the disintegration of this cemenl is unexplainable excepl thai il was do1 clinkered properly. The boiling test, however, showed a good cement. (Table VI.) As the crushing strength tests of the Portlands show, there seemed to be no weakening due to being in the salt solution. Table VI. — Results of Boiling Test fob »'» Horns, after (>() Hours in Moist Chamber. Number. Appearance after s,;i Water Test. .4i Good, cracked plate. Came loose from plate and showed some expansion. Same asAj. Good. Came loose from plate, warped. Good. Good. Cracked. ,i 2 At... At... •••( Bi Sound. Bt " Bi... Wenl '" pii B, Sound. C\ Cracked and swollen. Ct Sound. Ca Ct... Also the strength of the Portlands seems to average higher than the Iron Ore cements. (Table VII.) Five cubes of each cement of Series (' were then placed into the cylinder with a set of Portland cubes made at the Bame time. These were kept under pressure for 8 days. The results of this series were quite different as 4 of the 5 cemenl cubes were badly cracked and had begun to swell. C 2 , C», and & showed no signs of disintegration, but (\ was cracked and swollen badly. This cement, as the A Series, did not stand the boiling test and such an action would be expected from it under the extreme condi- tions in the pressure cylinder. The crushing strengths of Ct, C 3 , and d averaged lower than the H Series, d was BO BOfl that disintegration had evidently set in. 14 Williams on Iron Ore Cement. Table VII. — Crushing Strength of Cements. No. P: Cross-sect ional Area, sq. in. Crushing Strength. Total lb. Lb. per sq. in. Average, lb. per sq. in. 1.08 7680 0.975 4780 1.06 6650 1.045 5650 1 . 105 7750 Pi = Portlands in fresh water 3 weeks. 7100 4900 6280 4910 7020 p. 2 =Portland cement in fresh water 4 weeks. 0.97 7850 8700 0.95 6620 6970 . 97 7730 7960 P=pressure with Series B of the Iron Ore Cements. 6042 7876 p 0.97 5420 5590 1.25 IM.II 3S90* 1.025 7650 7470 (i as 7330 7470 1.01 7200 7150 6920 Iron Ore Cement in salt solution under pressure cylinder i days. Bi 1 .035 5810 5620 1 n7.-, 6720 6250 1.035 :, 1 _'i i 4915 1.06 4740 1460 1 .045 5200 I SCO 5241 B» 1 . 105 717H 6500 1.02 6620 6000 1 1 155 7500 7 H in 1.115 8430 7550 1.125 6680 5930 6616 B 1 09 1480 11 2D 1 .075 5180 1820 I in 5000 1540 1 06 6610 6240 1.12 6000 5350 5014 C 2 1.025 4200 4080 1.03 5400 4360 1 . 025 6320 6660 1.1 5N.-)() 5310 1.04 4850 4660 4914 C 3 1.05 2280 2190 0.97 1580 1660 1.1 2640 2400 1.00 ls.'H 1880 1.01 2500 2480 2110 C, 1.07 5220 3000 1.07 6630 6150* 1 in, 3630 3330 1.07 5140 4800 1.04 4050 3900 5757 Port Ian ds in Cylinder 7 day s with Series C. P 0.99 3000 3030 P 0.97 6720 6930* Only unaffected Portland cement cube. * Signifies not calculated in average. Willi IMS ON [RON ( >ki: I 1 \n \ i . |."> ( "om i.i BION8. As tlu 1 time for this investigation was limited, further work could not be done, and the conclusions which may I"' drawn from these results are limited. This much may be -aid. however: 1. The amount of lime or silicate of lime ought to be less when l-Vjt); alone is used in place of Ab<> ; . a- the lowesl ratio of Series .1 5.1 was the only one which stood the boiling test. Series H showed that the limits gave lined cements throughout, neglecting Bj which must have disintegrated due to some other cause. Series c showed that the lime and silica required increased as the lower ratio f>.44 disintegrated and the higher ratios were good. To sum this up, when all iron is used the R_< h : Si< L ratio should he below 5.1; when 0.17.") to 0.2 mols. \U > i- used with 0.825 to 0.8 mols. of Fe,( ) :i the ratios li,. between 5.1 to 7.22. If O.'M) to 0.4 mols. of Ab< ).. ; the ratio must he 5.8 or greater. This is hut a suggestion and will require further experimenting to show it definitely. 2. That cements with large amounts of Fe^ h will stand saline solutions better than cements containing Ab< b was shown in the test of Series (' where the Portlands were actually disinte- grated and the iron cements stood the same test. 3. The results seem to surest that if the amount of lime was reduced lower than 2.8 CaO in Bleininger's formula, better strength could be obtained. There was found in the bottom of the vessel, after each trial in the cylinder, a heavy muddy deposit which was principally hydrated lime and which appeared to have been leached from the cubes. This reduction of the amount of lime may not need to be as much as the results suggest if the raw- materials were clay and limestone in place of pure whiting, Al 2 (OH) 6 and flint. All of the iron cements would have stood the tots better if they had been allowed to stand in the atmosphere and age, thus giving the lime time to become calcium carbonate. The Portland cement, which these cement- were tested against, was one of the besl cement- on the market. It tested as follows: Initial set. 3 hr.; final set 1 \ hi.; tensile Strength of neat cement after seven days, 679 lb.; after 28 days. 771 lb.: and it> crushing strength is shown in the table-. This cement had also aged several months in the laboratory and was in the best of condition 16 Williams on Iron Ore Cement. to stand accelerated tests. The percent of lime given by Mr. William Michaelis is 63.5 per cent with a small amount of magnesia, MgO, 1.5 per cent. The cements made for this thesis are all above 66 per cent, this is only another evidence that these con- clusions are correct and the following formula is suggested as the center of a series of cements for further experimenting: 4(2.8 CaO,Si0 2 ) 0.8 (2 CaO, Fe 2 3 ) 0.2 (2 CaO, A1 2 3 ). from this vary both the amount of Si0 2 and CaO. Bibliography. William Michaelis, Jr., Engineering Xeirs, Vol. .58, pp. 04.5-646. Charles J. Potter, Journal Society Chemical Industry, Vol. 28. Newberry, Journal Society Chemical Industry, Vol. 1G, No. 11. A. Meyer, Ckemisches Central Blatt, Vol. 73, p. 1369. A. Spencer and E. C. Eckel, Patent No. 912,266, U. S. Karl Zulkowski, Chemische Industrie, 1901. A. W. Thackara, U. S. Consular Reports, June, 1908. Iron Ore Cement, The P. C. Co. of Hemmoor, Hamburg, Germany. Lamine, he Ciment, 1901, pp. Ill, 691, 81. Dr. Michaelis, Tom Industrie Zeitung, 1896, p. 838. Rebuffat, Tone Industrie Zeitung, 1901, p. 272. Le Chatelier, Le Ciment, 1901, pp. 31-32.