291 The Elastic Properties of Steel at Moderately High Temperatures .
By F. E. Rowett , B.A. ( Whitworth Scholar ) , Research Student in the University of Cambridge .
( Communicated by Prof. B. Hopkinson , F.R.S. Received February 6 , 1915 .
) In a recent paper the author gave the results of measurements of the elastic hysteresis of steel tubes , when subjected to torsional stress , within what is ordinarily regarded as the elastic limit .
Those measurements were made at ordinary room temperature , and the chief points established by them were , that the hysteresis of the hard-drawn tubes is much less than that for the same tubes after annealing , and that in the latter case , the loss of energy in the cycle of stress is independent of the speed of performance of the cycle , or , in other words , time is not a factor in the stress-strain relation .
The present research is the outcome of a suggestion by Prof. B. Hopkinson , that at a suitable temperature the hard-drawn tube , which contains a good deal of amorphous material , would begin to behave like a viscous fluid , that is , it would flow more or less freely when under stress , whereas at the same temperature , the annealed tube , being crystalline , though it might take a permanent set , could not flow , or would flow in a much less degree corresponding to the small amount of amorphous material in it .
These predictions have been verified .
At a temperature about 300 ' C. , the hard unannealed tube began to show properties similar to those of pitch at ordinary temperatures , or of glass at a temperature rather below the softening point .
It is still highly elastic under rapidly varying stress , but flows perceptibly when the stress is applied for a long time .
The energy dissipated in a cycle of stress now depends on the speed of the cycle .
If the cycle is performed in 5 seconds the area of the closed stress-strain loop is not much greater than at ordinary temperatures , and is less than that given by an annealed tube .
But if the cycle takes a quarter of an hour , the dissipation is increased four-fold .
On the other hand , in the annealed tube at 300 ' C. the energy lost per cycle of stress is still almost independent of the time .
At a higher temperature , for example at 540 ' C. , the hard-drawn tube flows rapidly and continues to flow for a long period , though at a diminishing rate , under a shear stress of less than 1 ton per square inch .
Moreover , again like pitch or glass , the steel at this temperature shows considerable elastic afterworking .
If the stress be suddenly removed the immediate elastic recovery is followed by a slow backward flow which persists for many minutes .
Some Mr. F. E. Rowett .
The Elastic Properties of flow and elastic afterworking were also observed in the annealed tube at this temperature , but both were much less than in the hard tube.* Apparatus .
' The apparatus was an improved form of the static apparatus described in \lt ; Roy .
Soc. Proc. , ' A , vol. 89 , p. 537 , and the method of obtaining the area and shape of the hysteresis loop has been explained there .
The experimental tube was rigidly fixed at one end to a much longer and stronger hard tube .
The stiff tube was rigidly fixed at one end , and the torque was transmitted to it through the experimental tube , the twist in the stiff tube providing a means of measuring the torque .
Fig. 1 gives a general view and details of the apparatus .
The compound tube G-H was rigidly fixed at G- , and the end H securely fixed to lever L , which worked between adjustable stops E and F. By movement of this lever any desired stress could be applied .
Sensitive micrometer levels were fixed to the tubes in the positions A , B , C , D , and at Mi and M2 small concave mirrors were arranged in conjunction with lamp and scales Si and S2 .
The weights W exactly balance the weight of lever L. The mirrors in conjunction with scales supply the means of measuring extreme ranges of strain and thus of stress .
The tube was heated by passing an alternating electric current through the tube itself , and the temperature of the tube was ascertained by three thermocouples in conjunction with galvanometer J and scale .
The current was led into and from the tube by means of small arms dipping into mercury baths , a mode of connection which did not exert appreciable force on the tubes .
It was found by experiment that by taking a tube sufficiently long , a fairly uniform temperature could be maintained along the central portion of the experimental tube .
Three thermo-couples were brazed to the tube at points G inches apart , and these showed a variation of 3 ' or 4 ' C. at 500 ' C. , the ends being below the centre in temperature .
The mirrors were set 8 inches apart , so that , for the length under observation , the temperature has been taken as that given by the central thermo-couple .
The thermo-couples were calibrated by observing the melting points of pure metals , and temperatures given may be taken as true to within 2 ' C. The heating current was taken from a transformer , the secondary circuit consisting of four coils arranged in parallel , each sending its current through the tube .
The temperature was readily regulated by external resistances in * Hopkinson and Rogers found considerable elastic afterworking in steel at 600 ' C. and upwards after applying and removing a load of 1^ tons per square inch ( ' Roy .
Soc. Proc. , ' A , ^ol .
76 , p. 419 ( 1905 ) ) .
Steel at Moderately High Temperatures .
the primary circuit , a resistance slide giving a delicate means of adjusting the current to give any desired temperature .
An electromotive force ranging from 20 to 200 volts by steps of 20 volts could be applied to the primary circuit , which together with adjustable resistances above mentioned allowed any temperature between that of the room and 600 ' C. being maintained for any period of time .
294 Mr. F. E. Rowett .
Th Elastic Properties of To obtain a uniform temperature along the tube all the heat was allowed to radiate freely to the atmosphere , no lagging being used .
The mirrors were protected from radiation by reflectors , and heat was prevented from conduction to them by supporting them at the ends of glass tubing as shown in fig. 1 .
The steel tubes had the following chemical composition:\#151 ; Carbon 017 percent .
; manganese , 0'24 per cent. ; sulphur , 0'02 per cent. ; phosphorus a trace , the remainder being iron .
The tubes as supplied by the makers were hard as the result of the drawing .
The elastic limit in torsion in this state was 13'8 tons per square inch shear stress in either direction , giving an elastic range of 27'6 tons per square inch .
By annealing at 800 ' C. for 15 minutes and cooling in an electric furnace , the effect of the drawing was removed .
The tubes were then quite soft and ductile and had a well marked yield point in torsion of 5'58 tons per square inch shear stress , giving an elastic range of 11T6 tons per square inch .
Hysteresis Experiments .
A hard-drawn tube was introduced into the apparatus and subjected to a stress range of 8-32 tons per square inch ( + -T16 tons per square inch ) ; 200 cycles between these limits were performed to bring the tube into a cyclic state , and the stress was reduced approximately to zero by bringing the lever L into the central position .
The micrometers were then adjusted to make the levels A and B both read zero .
The load was then applied and removed , thus completing another half-cycle of stress , and the stress again adjusted by movement of lever L until the levels A and B again read zero .
The effect of this procedure is to bring the stress in the thin tube to exactly the same small value that it had before the half-cycle was performed , but this condition has been approached from the opposite direction and the strain is slightly different by reason of the elastic hysteresis .
This change of strain is shown by a small difference in the movements of the levels C and D which have occurred as the consequence of the half-cycle of stress .
The following results were obtained , for the above stress range:\#151 ; Movement of C 0'3 , movement of I ) 24 , and the difference , T8 , is a measure of the elastic hysteresis .
The tube was then heated to 100 ' C. and 100 cycles were performed at this temperature .
The result was:\#151 ; Movement of C 0-3 , movement of 1 ) 2'2 , and the difference , 1-9 , is greater than before by a small but unmistakable amount .
The above cycles were performed at the rate of 5 seconds per cycle .
A slow cycle was next performed lasting in all 2 hours , the temperature of the tube being maintained at 100 ' C. The hysteresis was found to be exactly the samoas in the quick cycle , namely , T9 .
The experiment showed that the Steel at Moderately High Temperatures .
295 hysteresis increased slightly with temperature up to 100 ' C. , but was still independent of the time .
At temperatures in the neighbourhood of 200 ' the hysteresis was greater when the cycle was performed slowly , the material showing some signs of how .
Table I gives the results obtained with stress range throughout of 11-60 tons per square inch ( \#177 ; 5*8 tons ) , this stress being well within the elastic limit as ordinarily measured .
Table I ( Stress Range , 1160 tons per square inch ) .
Hard-drawn Tube .
Thermo-couple temperatures C. Movements of levels .
Difference = hysteresis .
1 .
2 .
3 .
C. D. Value cold 19 19 19 0*8 3*3 2-5 Cycle performed in 5 secs .
194 198 196 0*8 4T5 3*35 " " 16 mins .
195 199 198 0*9 4*70 3 *8 Cycle performed in 5 secs .
227 230 225 1 *0 4*8 3 *8 " , ,16 mins .
227 229 225 1 -1 6*7 5 *6 Cycle performed in 5 secs .
290 294 289 1 T 6*0 4*9 " " 16 mins .
290 294 291 2 0 23 *7* 21 *7 Value cold after above 19 19 19 0*8 3 *3 2*5 * Equivalent in level motion .
Micrometer head used for measurement .
The above Table shows that , at temperatures above 200 ' C. , time is an element in the stress-strain relation .
It also shows that the tube was not changed in structure appreciably by being so treated , since , when allowed to cool , the hysteresis returned to its normal value .
In the paper previously referred to it was pointed out that the hysteresis of a hard-drawn tube was increased about eight-fold by annealing at 800 ' C. and cooling in the furnace .
Flow Experiments .
At temperatures above 300 ' C. the flow became too great for measurement by the spirit levels and the strain was read direct by means of the mirrors .
The levels B and C were used for measuring stress , level B being adjusted to read zero when the tube was subjected to a definite stress , and level C being placed on the large tube near level B and adjusted to read zero when the tubes were free from stress .
In making an experiment the large tube was first released from stress by releasing the fixed end .
Level C was then adjusted to read zero , and the large tube fastened down at the fixed end .
A stress was then applied and its amount calculated from movements of mirrors .
Level B was adjusted to read zero with the stress Mr. F. E. Rowett .
The Elastic Properties of applied .
To keep the tube while flowing at a constant stress it was only necessary to adjust lever L so as to keep level B at zero .
The tube could be freed from stress by adjusting lever until level C read zero .
Throughout the tests described below the tube was subjected to 'a constant shear stress of 079 ton per square inch .
The application of the stress and its adjustment to the desired value took about half a minute , and a similar time was occupied in removing the stress and adjusting it to zero .
Test .
I.\#151 ; Tube Unannealed ( Stress = 0-79 ton per square inch ) .
Time from start .
Stress .
Temperature .
Scale readings .
Difference = strain .
M , .
Mo. mins .
0 1 Stress being 340 320 i J applied 1 299 *5 268 -25 11 *25 y Stress on 20 J .
1 Stress being 337 ' C. 297 *5 265 -25 12 *25 20k j removed 339 318 1 *0 x Stress off 25 J .
339 318 *25 0*75 At this temperature the strain which develops in 20 minutes after the application of the stress is about one-tenth of the initial elastic strain , and about one-quarter of this flow is recovered after removing the stress , leaving three-quarters as permanent set .
Time from Stress .
Temperature .
Scale readings .
Difference start .
Mj .
m2 .
= strain .
Test II .
mins .
0 k 2 4 6 8 10 15 20 20k 21 22 24 * j- Stress being applied y Stress on - Stress being removed | Stress off 394 ' C. 342 297 -5 296 6 295 -7 295 294 -5 294 293 292 336 -5 337 1 337 -5 337 -8 321 264 *5 263 261 *8 260 7 260 259*3 257 *7 256 312 *5 313-6 314 *3 314 9 12 *0 12*6 12 *9 13*3 13*5 13*7 14 *3 15 *0 3*0 2*5 2 *2 1 *9 Steel at Moderately High Temperatures .
Time from start .
Stress .
Temperature .
Scale readings .
Difference = strain .
mins .
0 20J Test III .
1 - Stress being applied 337 315 *5 \gt ; 291 *3 256 *5 290 *5 253 8 289 *3 250 *5 288 *2 248 *3 287 *2 246 *5 286 *3 244 *9 285 *5 243 *6 1 \gt ; Stress on 284 *8 242 *5 284 *2 241 *5 442 ' C. 283 239 *6 282 238 281 236 *6 280 T 235 *4 279 *2 234 *2 \lt ; * Stress being removed 278 *5 233 \lt ; 324 *2 292 325 293 *2 -Stress off 326 *5 295 *5 327 *5 297 328 297 *7 328 *2 298 *0 Test IV .
j* Stress being applied -Stress on Stress being removed Stress off 492 ' C. 329 *5 275 *5 271 *0 259 *5 238 *5 286 290 *5 293 *5 295 '5 298 298 *5 298*5 227 *5 220 209 *5 202 *25 197 192*5 189 186 183 178 173 *5 170 166 162 *5 159 230 237 241 *5 244 *5 248 *5 249 *75 13 *3 15 *2 17 *3 18 *4 19 *2 19 *9 20 *4 20 *8 21 *2 21 *9 22 *5 22 *9 23 *2 23 *5 10*7 10 *3 9*5 9*0 8 *8 8 *7 20 *5 23 *5 26 *25 28 29 *5 31 *0 32 *0 33 0 35 *0 36 *5 37 *5 40 *5 42 22 *5 18 *5 17 *75 Mr. F. E. Bowett , The Elastic Properties of Time from start .
mins .
0 20| Stress .
Temperature .
Scale readings .
Test Y. j- Stress being applied [ \gt ; Stress on j- Stress being removed \ [ \gt ; Stress off 542 ' C. M* Difference = strain .
348 *5 330 282 239 24*5 279 227 *5 33 *0 274 -5 216 40 *0 271 208 44 *5 268 201 48 *5 265 196 50 *5 263 191 53 -5 259 183 57 *5 255 *5 176 61 -0 253 '5 170 *5 64 *5 251 *5 166 67 *0 249 162 68 *5 245 155 71 *5 302 233 50*5 306 238 *5 49 309 243 47 310 *5 246 *5 45*5 311*25 248 44*75 312 *0 249 *5 44 *0 Test VI.\#151 ; The tube after Test Y was lowered in temperature to 491 ' C. , that is to approximately the temperature of Test IV , and records were taken of the flow under the same stress .
The following Table shows the results , and ( in the last column ) the amount Test YI .
Scale readings .
Time from Stress .
Tempera- Difference From start .
ture .
= strain .
Test IV .
Mj .
m2 .
mins .
0 i 1* Being applied 36 *5 -38*5 i J 84 *0 25 *5 16 *5 3 6 *9 1 1 88 3L *2 18 -2 20 *3 2 90 35 *2 20 *2 23 *3 4 93 41 23 27 *7 5 8 \gt ; Stress on 491 ' C. 94 96 *2 43 47 *6 24 26 *4 28 *2 32 *6 11 98 *5 51 *7 28 *2 34*3 13 100 54*5 29 *5 36 *6 15 101 *8 57 *5 30 *7 37 *8 20 * 105 *5 63 *5 33 *0 41 *5 - - - *\#151 ; 1 1 Steel at Moderately High Temperatures .
of flow observed at corresponding times in Test IV when the temperature was the same , correction being made for the 1 ' C. difference in temperature .
The difference is probably due to the annealing effect of Test V ( during which the tube was kept at 542 ' C. for half-an-hour ) , which would make the metal more crystalline and less liable to flow .
h to / CURVE I. UNANNEALED TUBES Stress 79 tens persq.in .
/ \lt ; o* 4/ / to / / / 'o* \amp ; \ .'\V .\#171 ; * ' ^:oo0- .1- \#166 ; I / / / w " 3^Te^P-4A2 / hi / " ' ' ' ' 7 -/ ' 20 1 / / / / ^ I/ / ?
' 16^ / f / 12j^ - ---------------2.Temp .
394 ' l \ I r 8r L i v o\lt ; - 1 .
Temp. 337 C. I [ \ I I l I .-I II II 'IP '|k *1 !
i L 10 12 14 16 Time in Minutes .
V. VOL. XCI.\#151 ; A. Mr. F. E. Rowett .
The Elastic Properties Test VII.\#151 ; An experiment was performed at 600 ' C. upon the same tube as Test YI and the following values were obtained :\#151 ; Test VII .
Time from start .
mins .
20^ Stress .
Stress on 0*79 ton per square inch Stress off Strain .
Temperature .
600 ' C. Experiments with Annealed Tube .
The tube was annealed by heating it to 800 ' C. in an electric furnace fox 15 minutes and allowing it to cool in the furnace .
The stress for annealed tubes was the same as that for unannealed tubes , namely 0-79 ton per square inch .
Time from start .
Stress .
Temperature .
Scale readings .
M2 .
Difference = strain .
mins .
0 20 \#163 ; Test I. j- Being applied 17*5 54 *5 -21 *5 27 y Stress on 55 28 J j- Being removed 401 ' C. 55 *5 18 *5 28 *8 -19 *7 | Stress off 18 *4 -19-9 Test II .
j- Being applied j- Stress on | Being removed Stress off 451= C. 18*0 56 *5 59 *0 61 T 22 *6 22 *5 -20 *5 37 *3 -13 *6 -13-9 11 *5 12*0 12 *3 0*8 0*7 12 *4 13*5 14 *7 2*1 i 0 h Steel at Moderately High Temperatures .
Time from start .
Stress .
Temperature .
Scale readings .
Difference = strain .
Mj .
Mo. F Test III .
mins .
o 18*5 -21 \#163 ; j- Being applied 59 -5 33 5 13 *5 10 ?
Stress on 62 *8 39 *3 16*0 20 J 499 ' C. 65 *5 43 *9 17 *9 20\#163 ; - Being removed 24 *5 \#151 ; 10 1 4*9 1 t Stress off ' 25 J 23 *9 -11 2 4*4 Test \gt ; HH 0 34 *5 -5'7 ?
Being applied i 75 *6 50 *6 15*2 2 77 *5 54 *5 17 2 5 80 *0 59 *3 19 *5 8 - Stress on 81 *7 62 *6 21 1 10 r are \o n 82 *6 64 *2 21 *8 15 ooU O. 84 *6 67 *7 23 *3 20 i 86 *3 70 *6 24 *5 ?
Stress being re20\#163 ; ; moved 45 *2 15 *5 10 *5 22 42 *5 11 *0 8*7 23 * Stress off 40 *8 8*5 7*9 25 38 *8 6*0 7*4 CURVE n. ANNEALED TUBES Stress *79 tons per sq .
in .
1.Temp .
40TC .
Time in Minutes .
\#166 ; ciin .
302 Elastic Properties of Steel at Moderately High Temperatures .
In the following Table the flow and recovery of the annealed ( S ) and nnannealed tubes ( H ) at the same temperature are compared .
The figures are corrected by interpolation from the results given in the above Tables to reduce them to the same temperatures .
\#151 ; Unannealed tube ( H ) .
Annealed tube ( S ) .
Temperature 500 ' C. 500 ' C. Probable elastic strain* 13 *3 13*3 Strain after J minute 19 *7 13 *8 Plow in 20 minutes 33*4 4-7 Recovery in 4| minutes 6*5 0*55 Stress 0 *79 ton per square inch 0 *79 ton per square inch * The elastic strain is that corresponding to an instantaneous application and removal of the load , and is calculated on the assumption that the elastic modulus falls 5 per cent , per 100 ' C. \lt ; o curve m. Relation between Flow in UNANNEALED \amp ; ANNEALED TUBES .
XX s ^ / p Stress *79 tons per sq .
in .
Temp. 500'C . .
\#151 ; ---------.TlSc *7ED TUBE -\#169 ; \#151 ; \#166 ; \#187 ; 2 4 6 8 10 12 14 16 18 Time in Minutes .
20 22 24 26 28 A Chemically Active Modification of Nitrogen .
The above work was carried out in the Engineering Laboratories of Jambridge University , and I wish to express my sincere thanks to Prof. B. Hopkinson for his many suggestions and inspiring interest .
A Chemically Active Modification of Nitrogen , produced by the Electric Discharge.\#151 ; VI.# By the Hon. B. J. Strutt , Sc. D. , F.R.S. , Professor of Physics , Imperial College , South Kensington .
( Received March 9 , 1915 .
) S 1 .
Effect of Catalysts in Promoting the Formation of Active Nitrogen .
There has been considerable controversy on the question of whether or not pure nitrogen would give the afterglow , which , as I have shown in the previous papers of this series , is associated with the presence of chemically active nitrogen .
E. P. Lewisf was disposed to think that the presence of oxygen or nitric oxide was essential , but in a much later paper J though still inclined to the same opinion , he states that the afterglow continually increased in intensity as the proportion of oxygen was reduced .
Warburg , in some experiments on the " luminous electric wind " in nitrogen at atmospheric pressure , S found that the intensity of the effect was much diminished by prolonged heating with sodium at 300 ' C. , though he expressly states that he was unable to get rid of it altogether .
It is not clear that the " luminous electric wind " ( which was not bright enough for adequate spectroscopic examination ) is the same phenomenon as the active nitrogen glow obtained at low pressures .
Experiments made later in Warburg 's laboratory by von Mosengeil|| on the active nitrogen glow at low pressure led to the conclusion that it was not diminished by heating the gas with sodium in a closed discharge vessel .
* I , 'Roy .
Soc. Proc. , ' A , vol. 85 , p. 219 ; II , ' Roy .
Soc. Proc. , ' A , vol. 86 , p. 56 ; III , 'Roy .
Soc. Proc. , ' A , vol. 86 , p. 262 ; IY , 'Roy .
Soc. Proc. , ' A , vol. 87 , p. 179 ; Y , ' Roy .
Soc. Proc. , ' A , vol. 88 , p. 539 .
+ ' Astrophys .
Journ. , ' vol. 12 , p. 8 ( 1900 ) ; ' Ann. der Phys. , ' vol. 2 , p. 249 ( 1900 ) ; and also 'Astrophys .
Journ. , ' vol. 20 , p. 49 ( 1904 ) .
t 'Phil .
Mag. , ' June , 1913 , p. 326 .
S ' Ann. der Phys. ' [ 4 ] , vol. 10 , p. 180 ( 1903 ) .
II ' Ann. der Phys. ' [ 4 ] , vol. 20 , p. 833 ( 1906 ) .

