,D-R125 IN GAS ENVIRONNENTSCU) JOHNS 1I HOPKINS UNIY .

*-.-.THE JOHNS HOPKINS UNIVERSITYAPPUED PHYSICS LABORATORYLAUREL.MARYLAND2. SUMMARYto 4500 F with incident gas flow rates from 1.2 to 22m/s. Air, oxygen (O2), carbon dioxide (C0 2), andcarbon monoxide (CO) were used as reagents. Sometests were run with "dry" reagents whose watercontents were typical of compressed gases. The watercontents of the compressed gases were not measured;however, on the basis of information from thesupplier, typical water contents at 2000 psia are 17ppm for air, 8 ppm for 02, 128 ppm for C0 2 , and 2ppm for CO. For some tests, water vapor was addedin known amounts. Also, mixtures of nitrogen (N2 ),02, CO 2, CO, and water vapor were used to simulatethe products of combustion of Shelldyne-H fuelburned at equivalence ratios (ER's) ranging fromapproximately 0.3 to 1.0.A summary of test conditions and results for the 40tests reported herein is presented in Tables 1 and 2.The results agree well with data in the literature forair and CO 2. CO2 and water both react with the ATJgraphite surface, even at temperatures as low as2000 F.The apparatus developed in late 1979 in the APLCombustion Research Laboratory' was modified forthe current series of tests. Figure I is a block diagramof the apparatus as it was at the end of the tests; Figs.2 through 7 are photographs of various parts of thetest setup and ancillary apparatus.Specimen temperatures up to 4500 F areachievable with the electrically heated graphiteelectrode. Specimens of ATJ graphite, approximately 0.13 in. (3.3 mm) in diameter and 0.07 in.(1.78 mm) thick (Fig. 5), are placed in a graphiteguard ring; the assembly is put into a shallowcylindrical recess at the tip of the heater and heatedby conductive and radiative energy transfer from theheater. Specimens can be heated this way to 4000 F;with less reliability, they can be heated to 45000F.Reagent gases exit from the nozzle, which is normalto the flat surface of the specimen (Figs. I and 3).The numbers in the triangles in Fig. 3 correspond tothe numbers identified in Fig. 1.Tests were run at temperatures ranging from 2000110.

THE JOHNS HOPKINS UNIVERSITYAPPLIED PHYSICS LABORATORYLAUREL.MARYLAND0d)30CLL.4-E 0)CL 0.-,00-Z--tsacm u. , I704-CL mm m 0m0.0.Lua,.4)EUULaw a ECZ0))-- 40(A 0(U %-plA2(U(-2i9--0E0.)C--N00(E00 E00 -0L

THE JOHNS HOPKINS UNIVERSITYAPPLIED PHYSICS LABORATORYLAUREL MARYLAND-multimeters0IFiure-Iret0raphte blaton8(forthermocouples)ppartus Pyrx coer n plce)cyind

THE JOHNS HOPKINS UNIVERSITY-APPLIED PHYSICS LABORATORYLAURELMARYLAND4).073IA.ks'L4-)

F.Z-**Z,'*%THE JOHNS HOPKINS UNIVERSITYAPPLIED PHYSICS LABORATORYLAUREL.MARYLANDLuer fittingmixing chamberThermocoupleFigure 4 - Vaporizer assembly.14

flw Jce IIs HQ"SUNIVEFI ,ry1APPUED PHySICSLABORLATORYLAUREL. MARyI.ANDNIIF gre610- Thrm co p8s mlfrfze-n----

TIE JOHNS HOPINS UNIVERSITYAPPLIED PHYSICS LABORATORYLAUREL,MARYLAND4447:4AFigure 7 - Device for measuring specimen thickness.16

TMOHNS HOPKNS UNIVERSWYAPPUED PHYSICS LABORATORYLAUREL. MARYLANDTable 1 - ATJ graphite ablation test summary (1981 tests).TestNo.ATJTemp.TmF)GasWater(')(VOL %)Gas Velocity (mis)Nominal(b) ionRate 6.619.9300.1791.020.9017

THEJOHN HOPKINS UNIVERSITYAPPLIED PH{YSICS LABORATORYLAUEL. MARYLANOTable IA Timassatr(iTet em.N. (F(cont'd)Gas Velocity (mis)TestTransfer(s)91RecessionRate ed0.38212000Air223500ER 0.3(c)2022.5400.1651.043500ER 0.3e)2022.5400.1721.081.02242000ER 0.31c)11.181200.1030.140.10253500ER 0.3(e)12.01770.1500.240.17263500ER 0.3(c12.011800.1530.250.21272000ER 0.6(e)11.183000.08020.110.097283500ER 0.6 9302000ER].1Ole'11.164500.0210.0280.017313500ER- 1 CO,Dry12.481200.1170.240.19353500CO6.5%12.484200 30.1560.330.18393500ER12.483000.0890.150.124000ER 1.0112.482400.0810.140.11g23*40Ga--0.6(e' 21.0"'(ai Water content indicated is percent by volume of total (gas plus water). "Air"' and(CO, dry- contain :400 ppm water ( 0.040% by volume).N Based on room-temperature exit gas from nozzle.IdIdiof the test.There may have been less water vapor in the last partWater content is uncertain.simulated as products (it combustion (N,. 0. CO., CO. 1-1O) for Shelldyne-H fuel;see Table 3.180.90

THE JOHNS HOPKINS UNIVERSITYAPPLIED PHYSICS LABORATORYLAURIEL MARYLANDTable 2-Summary of conditions under which tests 1through 40 were run.Nominal Gas Velocity* for Sample TemperatureGas Conditions200VFImis 16.6m/stAir 400 ppm H2 04.5% H 206.5%H 2012Simulated Shelldyne-airER 0.3Im/sIm/s4ImIs 16.6m/s3211314II2425,26ER 0.62728,29ER 1.03031,39026.5% H20*3500"F2500"F 29,104500"FIm/s1522,234036CO2 400 ppm H 206.5% H 2 0343233Co6.5% H 2 035Argon6.5% H 2 037'Based on room-temperature exit gas from nozzle.tThere may have been less water vapor in last part of test.ttWater content uncertain.1938

THE "ONSHOPKINS UNIVERSITYAPPLIED PHYSICS LABORATORYLAUREL,MARYLAND3. TEST APPARATUS DESIGN CONSIDERATIONS*One objective of this effort was to measuregraphite ablation rates under conditions for whichdata were not available in the literature or in whichthere is special interest relating to the study of hightemperature thermochemical reactions of graphite.Another objective was to develop an apparatus andtechniques that - in addition to their purpose in thisseries of tests - could aid planned future investigations using Raman scattering to measurespecies concentrations in the reaction zones near thesurface of heated graphite.The graphite specimen, ring, and shield assembly(Fig. 5) presents a continuous surface to the gas sothat the pellet (specimen) will be uniformly ablatedwithout undesired edge effects. The assembly rests onan electric resistance heater consisting of a 0.065 in.(1.65 mm) thick graphite sheet, 0.75 in. (1.91 cm)wide and 2.5 in. (6.35 cm) long, with a saw cut alongmost of its length to form a U-shaped resistor. Thetop of the heater (Fig. 5) is counterbored to a depthof approximately 0.030 in. (0.76 mm) to provide apositive support for the specimen and to help keep itfrom being blown away by the reagent gases. Watercooled "jaws" provide the electric connections andsupport the heater electrode. One set of jaws ismounted on coiled stainless steel tubing (Fig. 3) toallow one leg of the heater to move freely. Thisprovides minimal resistance to movement resultingfrom unequal expansion of the two legs of the heater;the legs can expand without breaking as they woulddo if they were supported rigidly.A nozzle, 0.5 in. (12.7 mm) ID, is used to reducethe possibility of undesirable effects of contamination of the reagents as a result of the entrainment of gases present in the enclosure anddiffusion of the ambient environments into the gasstream. It also ensures stagnation flow conditionsacross the entire surface of the 0.22 in. (5.6 mm)diameter specimen-guard ring assembly. The exitplane of the nozzle is 0.43 in. (II mm) from thesurface of the specimen.To ensure that the gases are not substantiallybelow room temperature as a result of adiabaticexpansion (as the gases are released from the compressed gas cylinders), the gas supply line is fittedwith a simple heat exchanger made of coiled coppertubing immersed in a container of water at roomtemperature (Fig. I).A simple apparatus has been made for injectingwater vapor into the gas stream. A well-controlledflow of water vapor can be introduced into the testgas by means of a soldering iron heating element(Fig. 4) to supply heat and a commercial syringepump (Fig. 2) to inject water at the required ratethrough small-gauge stainless steel tubing into thevaporizer. However, the output capacity is limited;consequently, tests with high water content (5 to10%) were limited to 1 m/s velocity. Also, because ofthe limited water vaporization capacity, an accuratesimulation of the Shelldyne-air combustion productsfor the ER I condition (which requires 9.2%water) could only be achieved for a flow velocity of Im/s. To prevent condensation from occurring in theinlet piping and nozzle assembly, these parts werewrapped with heating tape.A data acquisition system recorded all datameasured during the tests. Eight channels of aHoneywell Model 1858 CRT Visicorder were used torecord the analog voltages indicated in Fig. I. Adetailed description of each channel is provided inAppendix A.Accurate measurement of the ablation is difficultbecause a lower-density layer is formed near thesurface when ablation first starts. This lower-densitylayer is evident in an electron microscope photographof a specimen's cross section (Fig. 8). The thicknessof the layer was determined to be about 0.008 in.A'IFigure 8 - Scanning electron microscope image of ATJpellet after test.2(

ThJOHNS HOPKINS UNIERSITYAPPLIED PHYSICS LABORATORYLAUREL.MARYLANO.along the stem of the dial indicator is attached to asmall ball bearing epoxied to the indicator stem. Bymeans of an ohmmeter, very light contact with thespecimen surface can be detected easily. Specimentemperatures were measured using the disappearingfilament pyrometer described in Appendix B.(0.20 mm), and it was easily compressed when amicrometer measurement was attempted. Toovercome this problem and to obtain an accuratemeasurment of the pellet thickness, the special deviceshown in Fig. 7 was made and was found to worksatisfactorily. The wire that can be seen emerging21

TH1EAONS HOPKINS UNIMOIRISTYAPPLIED PHYSICS LABORATORYLAUEL. MAMf4. TEST PROCEDUREThe general test procedure is as follows:I!. The pellet sample is weighed on an analyticalbalance and is measured with a micrometer.Then it is installed in the guard ring and electricheater.2. Power to the heater is supplied by a 5 kVAtransformer and is manually controlled by aVariac.3. The desired amount of each reagent gas ismetered through a choking orifice. The gasesare mixed in a chamber, and flow of the gases isbegun and stopped by solenoid valves.Unreacted gases and products of combustionleave the bell jar through a l-in.-diameter holein the baseplate directly beneath the tip of theheater and pass into the l.aboratory's ventsystem. The solenoid valves are controlled by aspecimen is conditioned. Meanwhile, thereagent gases flow through the orifices into themixing chamber but are vented from thechamber by valve 3 and are prevented fromentering the nozzle by valve 2 (see Fig. I). Thegases are vented to keep the mixing chambernear atmospheric pressure, thereby minimizingpressure transients at the orifice ontlets whenthe solenoid valves operate.4. When the specimen reaches the desired temperature (as determined by the operator viewingthe disappearing-filamentitthroughpyrometer), the operator throws a toggleswitch, which simultaneously terminates theflow of argon through the nozzle, closes thevent valve, opens solenoid valve 2 between thesingic manually operated switch that is connected to a timer, which is set for the desiredtest duratior. When the specimen is in place, thetest chamber is purged of air by argon admittedthrough a manually operated valve at the top ofmixing chamber and the nozzle, and starts theclock. These conditions persist for the durationof the test (during which the operator holds thetemperature of the specimen constant bycontrolling the Vafiac). At the end of the test,the chamber. This valve is closed manually, andargon is then passed through the nozzle onto thespecimen. In this locally inert environment, thespecimen is brought up to temperature and heldthere while the system stabilizes and thethe heater power is automatically turned off bythe timer, and the solenoid valve-, are switchedto their pretest conditions.5. The specimen is removed, weighed, andmeasured.i22

THE JOHNS HOKINS UNIVERSItYAPPUED PHYSICS LABORATORYLAUREL.MARYLAN1.5. DATA REDUCTIONRecession rates (U) have been measured as functions of surface temperature, impinging gas composition, and gas velocity. Eight tests wereduplications of previously run tests (Table 2); themeasured recession rates for those cases differ fromeach other by an average of 60 and at most by 15%.The average recession rate based on measuredthickness change is*(I)These calculated values are in good agreement withexperimental data (Fig. 9), indicating that the porouslayer remains nearly constant.When total recession is large compared with the 8mil porous layer, -Vm/), approaches unity. Mostablation tests reported in the literature use largesamples and have large total recessions; Y m/., approaches 1 for those cases. Since thicknessmeasurements are difficult and the mass loss methodis more representative of the amount of materialwhere y is thickness, t is time, and subscripts 0 andfrates" presented in the rest ofablated,theare.Y"recessionthisreportn.yt- (Yo - o)e.designate initial and final conditions.thsrptAn "equivalent recession rateAromthe mass loss:S(omf)yo/[mo(t-(Yin)-to))The test results plotted in Fig. 10 cover a large spanis computedof test conditions. The effect of gas velocity can beremoved from these data if the recession rates arenormalized by dividing .m by hi/p, where hi is the(2)local enthalpy-based heat transfer coefficient. The.The ratio j,/9, is plotted in Fig. 8 as a function oftotal recession based on mass loss (y,.), wherey. .m(tf - to). Figure 9 indicates that .)rm isconsistently greater than . , because of thedevelopment of an 8 mil (0.2 mm) deep, 50% void,porous region at the surface (Fig. 8). Assuming thatthe depth of this layer remains constant as recessioncontinues, then10m/s0 400 ppm4.6%6.5%0.30.30.603 400 ppm3.0%5.9%1.01yi/ y,/(y,,- 0.008 x 0.5)(3).a Measured dataCalculated by Eq. 3. assuming a porous1.9"llayerI 8 mils thickwithI 50% voidsI[I, 400 ppmCO 26.5%CO6.5%02Argon6.5%6.5%A000 9.2%Co210 m/s 16.6 m/s 20 m/sIVc"0*081.2I1--1.8Nominal velocity% HO2ER-11.71.6 1.51.41.3-2 0.6 -a a" U1.2 1.1ia.-.0*1020I1II3040500.42E'oa60LUEquivalent recession length, Yn, computedfrom mass loss rate, i m (mis)I02000Figure 9 - Ratio of equivalent ATJ graphite recession ratemeasured by mass loss method to that measured bythickness change.I30004000Surface temperature ( F)5000Figure 10 - Recession rate of ATJ graphite samples.23

THE JOHNS HOPKINS UNIVERSITYAPPUED PHYSICS LABORATORYLAUREL,MARYLANDnew quantityparameter, Y h,3is known asthe mass transfer8o:-h, 'h,(4)h,700Owherep density of virgin sample, andm mass loss rate per unit area.600 -In our experiment, h, is calculated from a formulation' for flow from a nozzle impinging on aninfinite flat plate:500h, 0.73b(Bg p ) 5(j p /p.)05/Pr 0 w(5)Nominal velocity1 mr/sILTwhereb blowingcorrection[ln (I 0')J/1(3'factor6.1 rn/sS400 B velocity gradient 1.5 Uax/d 2(s 'g gravitational constant (32.2 Ibm-300ft/lb f-s 2 )20-Pr The Prandtl number (cpj/k) ,ti. and , freestream and wall gas viscosities(I bf-s/ft 2) ,p.iooand p,, freestream and wall gas densities(lbm/ft')the velocity(ft/s) ,leavingthe2000nozzle3000ATJ qraphite surface temperature4000(OF)Figure 11 - Measured gas temperature at nozzle exitplane.d the nozzle diameter (ft) , andx the distance from the nozzle to thespecimen (ft) .leaves the nozzle exit plane at room temperature. Theactual velocity (U.) is estimated by multiplying thenominal velocity by T / Tro,,. One shouldremember, however, that T.0 is difficult to measureaccurately because the sample and heater surfacesradiate heat to the thermocouple, thereby introducing error. Although the thermocouple is shielded,the shielding method allows some reflected radiationto reach the thermocouple. Consequently, measuredtemperatures are higher than the actual gas tem-The freestream temperature (T) of the gas as itleaves the nozzle is difficult to measure, and p., u.,and U are all calculated from it. Because of heatfrom the boron nitride nozzle, T. can be substantially above room temperature. Late in the testseries, a radiation-shielded thermocouple was inserted along the nozzle axis to measure the T. of thegas at the nozzle exit. Those measurements were usedas a calibration to deduce the gas temperature for allof the ablation tests and for any specimen temperature and gas velocity. Figure I I shows T. versusATJ graphite surface temperature for several"nominal" air flow velocities calculated frommeasured gas mass flow rates, assuming the gasL. L. Perini, Reciew of Graphite Ablation Theory and ExperimentalData, .HU/AP! ANSP-M-1 (Dec 1971).I A. Belov. "The effect of Turbulence on Heal-Exchange in aJet Meetingan Obstacle," translated from TeploiMassoperenos(1969), British l ibrary lending Division.24

THE JOHNS HCPINS UNMERSTYAPPLIED PHYSICS LABORATORYLAUREL.MARYLAND"610 to 70OF is to reduce hi by 15%.The blowing correction factor (b) accounts for theablating combustion gases increasing the boundarylayer thickness and reducing the heat transfercoefficient. In the present tests, 0' is usually less than0.2; therefore, the blowing

R. W. Newman and C. H. Hoshall " N00024-83-C-5301 9. PERFORMING ORGANIZATION NAME & ADDRESS 10. PROGRAM ELEMENT, PROJECT, TASK AREA & WORK UNIT NUMBERS The Johns Hopkins University Applied Physics Laboratory Johns Hopkins Road Task X8GI Laurel, MD 2070