US 2057661 A
Description (OCR text may contain errors)
J. D. LE VAN Oct. 24), 1936.
Filed Jan. 27, 1952 JAMES DZiV/W v y Patented Oct. I 20,
SEAL I James D. Le Van, Belmont. Mass, assignor, by
mesne assignments, to Raytheon Manutacturing Company, Newton, Mass, a corporation oi Delaware Application January 27, 1932, Serial No. 589,180
This invention relates to seals of the type in which a hermetic seal is established between a metallic electrical conductor and a body of glass.
One of the objects of my invention is to provide a seal between a solid lead-in conductor of relatively large size and a body of glass of such a kind that the temperature of the lead-in conductor may be rapidly raised to a high value without subjecting the glass to fracturing stresses.
The foregoing and other objects of my invention will be best understood from the following description or" an exemplification thereof, reference being had to the accompanying drawing,
Fig. l is across-sectional view of one end of an electric discharge tube illustrating one embodiment of my invention;
Fig. 2 is an enlarged cross-section taken along line 2--2 of Fig. 1;."and
Fig. 3 is an enlarged perspective view showing an elemental portion of the glass layer on the surface of the lead-in conductor.
In sealing electrical conductors through glass, much work has been done on the problem of obtaining seals of the proper strength under conditions in which the temperature of the leadin wire remains fairly constant or else changes very slowly. This is a condition which is usually obtainedonly in electrical devices of comparatively small power rating. When, however, we contemplate using seals on electrical devices of comparatively large power rating, it becomes very difiicult to obtain the above temperature conditions for the lead-in-conductor. It is often desirable to change the load on such high power devices very rapidly, sometimes changing from no load to full load in a few seconds. Under these conditions the temperature of the lead-in conductor fluctuates widely and rapidly. Ordinary glass seals having lead-in conductors'of a size sufiiciently large to carry the loads almost invariably crack under such load and temperature variations. Previous workers in this art have attempted to overcome this disadvantage in various ways. Small lead-in conductors embedded in large thicknesses of glass have been used. Under these conditions, while severe stresses may be set up in the glass close to the surface of the lead-in conductor, yet the large body of glass. in the outer portion of the seal has had suflicient strength to prevent any appreciable stretching of the inner portion of the glass, and therefore has prevented cracking thereof. which large currents are used, small lead-in conductors are undesirable. In order to utilize large lead-in conductors, use has been made of hollow tubular metal conductors having a thin edge sealed into the glass member. This edge 0 is so thin that under the strains developed, due to heating by electrical currents, the metal in the thin edge reaches its yield point before any great strain is placed upon the glass of the seal. Thus under these strains, the metal in the thin edge actually flows. However, in such seals, the deformation to' which the metal is continually subjected ultimately results in the crystallization thereof, whereupon the metal becomes brittle and cracks. I have discovered that a large solid metal lead-in conductor may be sealed in a glass body and the temperature of this conductor may be varied over wide limits within very few seconds without the glass of the seal cracking if the thickness of the layer of glass in contact with the lead-in conductor is made small in comparison with the size of the lead-in conductor. I have further discovered that the thickness of this layer cannotbe greater than a definite upper limit, depending upon the materials used and 1 the maximum temperature variations to which the lead-in conductor is subjected.
While the glass layer actually in contact with. the metallic lead-in conductor has a tendency to crack, a similar tendency exists in the glass immediately adjacent the pointwhere the contact between the glass layer and the lead-in conductor ceases. This appears to be due to the fact that under rapidly changing temperature conditions, a comparatively large temperature difierence exists between the glass in the seal itself and the glass immediately adjacent the point at which contact between the lead-in conductor and the seal ceases. Th a large temperature gradient exists in the glass at this point, which large temperature gradient sets up excessive strains in the glass. I have eliminated this difliculty by providing means for distributing the temperature difierence between the glass in the seal and the rest of the body over a con- Obviously in large power devices in 2 a siderable length of glass, whereby the temperature gradient at every point in the glass is less than that at which fracturing strains-are developed.
Referring to Fig. 1, I is the wall of a glass vessel. This glass vessel may be, for example, any.
type of electron discharge device, such as, for example, an evacuated tube, a gas-filled rectifier, or the like. The vessel is provided with a reentrant stem 2 having a seal 3 at the inner end thereof. Through this seal passes a solid metallic conductor 4 which may be of any suitable metal. This conductor 4 is of a comparatively large diameter, which diameter is preferably in excess of one-tenth of an inch. Of course it will be understood that lead-in conductors of much larger diameter may be used, and also the conductor may be of a somewhat smaller size. The seal 3 comprises a thin layer 5 of glass fused around the conductor 4, whereby a hermetic seal is formed between the conductor and the layer 5. The conductor 4 may be used to conduct electricity to the interior of the glass vessel for any purpose whatsoever. example, in the drawing the conductor 4 leading to an anode 6. The device would also include a cathode, not shown, and would be connected in some.suitable circuit. For example, I have shown a wire 1 connected to the outer end of the leadin conductor 4. This wire leads 'to the secondary 8 of transformer 9 having a primary l0 connected to some source of alternating current. A
conductor ll leads from the other end of the secondarytl to a switch l2, while a third conductor l3 leads to a load device 14. The other terminal of the load device is connected by. means of the conductor I5 to the cathode, not shown.
With lead-in conductors of large diameters, I have found that the thickness of the glass layer 5 must be kept below a certain critical value for definite maximumtemperature variations of the lead-in conductor 4. I have thoroughly analyzed each of the factors upon which this thickness depends, and have discovered, in accordance with the following analysis, just what this thickness should be in each case. When the diameter of the lead-in conductor 4 is large compared to' the thickness of the glass layer 5, an elemental or unit section of the layer 5 in contact with the surface of the conductor 4 approximates a flat plate.
Referring to Fig. 3, I have illustrated such an elemental or unit section, and have shown this section substantially fiat. In order to analyze the problem, I designate the metal 4 as a body A and the glass layer 5 on its surface as a body B. The thickness of the layer B may be represented by X.
In order to determine the manner in which the at the center of said block, this center being at a distance from the surface of the metal body A.
In the table below is given a list of symbols representing constants bf the glass layer and the metal body A, which will be used in the following analysis.' The values of these constants for I have'illustrated, by way of 'a borosilicate glass and for the metal tungsten which I have utilized irLone embodiment of my invention are also given.
The degrees temperature specified are degrees centigrade:
I =therma1 conductivity for a centimeter cube=.0027 L sec. deg. cm. gm.
p =density=2.25 cm} c deg. gm. C=heat capacity=spv where v=volume of glass element =thermal' resistance for a centimeter For glass 1 cube= R=thermal resistance for 1 cm. cross section to a distance of i=2 K so that T4=Kt (1) where t=time in seconds Under these conditions it can be calculated that the maximum temperature difference becomes (TATB)max=K(Rc) (2) In the above Equation (2) RC is called the time constant of the glass, and gives the time at e which for a sudden application of a definite temperature to the surface of the glass body, the temperature of the glass body will have risen to or 63 per cent. of its final total rise. In this expression 5:2.71828.
Referring to Fig. 2, the metallic conductor 4 corresponds to body A and the glass seal or layer 5 to layer B, as in Fig. 3. v
In order to find the strains and stresses in the glass layer B, due to heating, we can assume that at some temperature To, the glass layer is free from strains. 7
If we now raise the temperature of A to some value TA, strains will be set up in the glass layer which can be expressed as a fractional stretch of the glass layer.
In increasing from To to TA, the metal A will undergo a fractional increase in circumference IA=(TA-T0) EA where Ea=coefiicient of expansion of A. Likewise the glass B in increasing from To to TB will undergo a fractional increase in circumference 1B.
where En=coeflicient of expansion of B. Then the fractional stretch DB of B is the difference between (3) and (4) which is s =thermal capacity per unit of mass= can use any final temperature Te such that the steady state stresses are less than the maximumin practice a= s and therefore- Ea=Bs+Y (6) or Y - Y.=E4-Es where Y is a constant. 7 Substituting in (5) and simplifying Da=Y(T4-To) +Es(T -Ts) ('1) Again referring to the glass. section under the area Z, we can substitute R and C in terms of fixed constants of B and the thickness X of the glass layer.
An examination of the equation yields very valuable information. The factor (Es-Es) (TA-T0) is a measure of the final or steady state stres conditions. The factor r: K I A B 21 however represents the transient stress condition since its value depends directly on K, the
rate of rise or the temperature Ts with respect. to time (see Equation (1)).
In applying the Equation (10) to a particular case we know the maximum permissible stress that the glass which we are using can stand. We
permissible stress and also a certain portion of the mammum permissible stress is left over to provide for the transient stresses. The maximum transient stress while proportional to the first power of the rate or change of the temperature Ts of the lead-in conductor is also dependent upon the second power of the thickness X of the glass seal. Thiis by keeping this thicmess small, we are enabled to use large rates of temperature change. In any case witha definite rate of temperature change, there is a definite upper limit to the thickness which the seal can assume without subjecting the glass to fracturing stresses.
As long as our approximation of fiat surfaces holds, we can see that the thickness of the glass layer'is independent of the size of the lead-in conductor. Thus the same thickness can be used for very large lead-ins as well as smaller ones. This approximation holds good as long as the thickness of the glass layer is small in comparison with the size of the lead-in conductor. However, with lead-ins which are so small that the thickness of the glass layer is not small as compared to the diameter of the lead-in, our approximation of flat surfaces no longer holds and the above analysis no longer gives an accurate picture of the actual conditions. In order to have this approximation hold, the radius of curvature of the surface should preferably be at least of approximately the same value as the thickness of the glass layer. However, it is still more advantageous to use even larger radii of curvature for said. surface.
'l'heaboveanalysiscanbeappliedtoaseai having any kind of metal and any kind of glass. In a particular instance in which I have used this seal, I utilize a conductor of tungsten about in diameter. The wall I consisted of a borosilicate glass, such as, for example, that known in the art as Nonex". The. inner part of the stem. 2, including the layer 5, was formed ofa glass which closely matched the coefllcient of expansion of tungsten. For this glass I used another borosilicate glass, which is known in the art as Corning 705 R". 01 course it will be understood that any suitable metal may be used for the conductor 4 and any suitable glass for the seal. We can now apply the analysis to the particular kind of glass and metal as specified above, and calculate the value of X for this particular example. The tensile strength P of the glass as taken from the table above is, of course, the
maximum permissible stress which can exist in the glass without fracturing it. Thus this value of P can be substituted in Equation 10 for St. (mu) -We can assume for the sake of simplicity To=0 C.
' which is a condition more adverse than is met with in practice. Therefore a solution on this as sumption will give a value of X which is well within safe limits.
Suppose we wish tooperate the lead-in conductor lat a temperature of 600 C., at which temperature it is at red heat, and raise it to this temperature in thirty seconds, we can readily see that these conditions are very extreme, and are such that they have not heretofore been successfully met in practice. From the above conditions our constants for Equation (10) become To=0. TA=600 K=20 stm=6oo The other constants can be taken from the table of constants given above. Suhstithting these values in Equation 10 and solving for X, we get X=.19 cm.
Time (T -T X constant Max.
Seconds 1 mm. 833 16. 66 C 2 mm. 3. 333 66. 66 C 5 mm. 20. 833 416. 66 C It is evident that since increases in thevalue of X give such large corresponding changes in time constants, temperature difference and consequently transient strains, we are sharply limited to the safe value of X as calculated for our glass seal thickness.
While in the specific solution which I have given by way of example above, the constants for tungsten and .a bore-silicate glass were used, yet
for practically any metal and glass seal the constants of the materials will be of the order or the constants used above. Therefore the order of thickness of the glass in the seal will be very close to that given by the specific solution above.
From thejoregoing analysis it will be seen that the radial thickness of the'glass wherever in contact with the conductor 4 must be less than a certain predetermined maximum. At the point l6 where the glass leaves contact with the surface of theconductor 4, there is a tendency for the radial thickness of the glass to be greater than the value specified unless special precautions are taken. By making the angle between the surface of the conductor 4 and the'surface of the glass at the point where contact ceases small, the radial thickness of the glass at this point can readily be kept within the requisite value. This angle should preferably be of the order of 10 to While the provision of a thin glass layer 5, wherever in contact with the conductor will prevent the cracking of the seal in this layer itself, the glass still has a tendency to crack somewhere in the region of the point i6 unless additional special precautions are taken. This is due to the fact that the temperature of the glass in the layer 5, due to' the thinness of this layer, varies substantially as does the temperature of the lead-in conductor 4. The temperature of the glass not in actual contact with the conductor 4 varies at a much lower rate, due. to the fact that its temperature is raised practically entirely by heat conducted through the glass from the layer 5. Thus a sharp temperature gradient would exist at pointlG when the temperature of the lead-in 4 varies at the rapid rate which I use.
It can be shown by an analysis similar to that given above that if at any time the temperature gradient in the glass at any point becomes greater than a certain value, fracturing strains would be set up in the glass at that point. In-order to decrease this temperature gradient, I provide means'to distribute the temperature difierence between the glass layer 5 and the restof the stem 2 'over a considerable length of glass. In the particular embodiment illustrated, I fill the space between the conductor 4 and the inside walls of the stem 2 with a finely-divided powder ll of good heat conductivity. This powder may be, for example, a powdered metallic substance, such as, for example, powdered aluminum, copper, tungsten, or the like. In order to retain the powdered material in place, I insert a packing l8 immediately above the powdered material l1. This packing is preferably constituted of a metallic foil, such as, aluminum foil, although a cement or sealing compound of some kind could also be used.
When current is led through the conductor 4. the temperature of said conductor rises, the tem- ,terial is also-in intimate contact with the conductor 4, heat is readily transmitted from each of these elements to said powder. This powder is packed suificiently densely so that heat is readily transferred therethrough. passes from the seal 3 and the conductor 4 is transmitted through the powder l6 tothe inner walls of the enlarged section of the stem 2. Thus The heat which 1 the temperature difl'erence between the walls of the stem 2 and the seal 3 is distributed over a substantial length of glass so that large temperature gradients do notexist at any point in said glass. It will be noted that the walls of the stem 2 are considerably thicker than that of the seal 3. However, the presence of the metal pow- "der I! in contact with these thicker walls does not introduce any excessive strains on the glass walls of the stem 2, inasmuch as the metallic powder I 6 readily yields under any expansions and contractions of the walls of the stem 2, due to temperature increases and decreases.
The presence of the metallic powder I I also eliminates another efiect' which exists in the vicinity of the point l6. It will be noted that at thispoint the glass and the metal in the conductor 4 are constrained to expand and contract at substantially the same rate, due to the close bondbetween the glass and the metal at this point. However, the glass walls of the stem 2 at I have constructed a large number of seals of the type described above, using a solid tungsten lead-in rod, the diameter of this rod being about two-tenths of an inch. In many. of these seals I have increased the temperature of the conductor by passing current therethrough from room temperature to red heat in a period of about thirty seconds. In each caseno fracture of the glass resulted. However, in similar seals without the use of the thin glass layer and the temperature-equalizing arrangement, practically all of the-seals cracked.
This invention is not limited to the particular details of construction, materials, and processes described, as many equivalents will suggest themselves to those skilled in the art. It is accordingly desired that the appended claims be given a broad interpretation commensurate with the.
scope of the invention within the art.
What is claimed is:
1. A glass to metal seal comprising a tubular glass member closed at one end, a metal rod sealed directly to the glass through said closed end, leaving a space between said conductor and the side walls of said tubular member, and a material of good heat conductivity placed in said tubular member in intimate contact with said seal and the adjacent side walls of said tubular member and of said conductor.
2. A glass to metal seal comprising a tubular glass member closed at one end, a metal rod sealed directlyto the glass through said closed end, leaving a, space between said conductor and the side walls of said tubular member, and. a nonrigid material of good heat conductivity placed in said tubular member in intimate contact with said seal and the adjacent side walls of said tubular member and of said conductor.
3. A glass to metal seal comprisinga rod of metal, a ring of glass completely surrounding said rod and adhering thereto throughout the circumference of said rod, the thickness of said than glass ring at every point thereon being not greater To=normal temperature of seal at which subv stantially no strains occur in said glass ring,
Tx=maximum steady state temperature which said seal reaches during operation,
MB'=Youngs modulus for said glass, S=thermal capacity per unit of mass of said glass,
, =density of said glass,
K=rate of temperature change of said metal rod of the order of magnitude of 20 C. per second or r'nore,
the upper edges of said rifig being extended to form a glass body surrounding said rod but spaced therefrom, the thickness of said tubular body at the point where its spacing from said rod starts also being of said order of magnitude.
4. A glass to metalseal according to claim 3 in which said metal rod is solid.
5. A glass to metal seal according to claim 3 in which the angle between the walls of said tubular member and the walls of said rod at the' point where the spacing between said walls starts is sufficiently acute to prevent an increase in the thickness of the glass at said point as measurea in a direction at right angles to the surface of said rod.
6. A glass to metal seal according to claim 3 in which means are provided for distributing the temperature diiference between the glass adhering to said metal rod and the rest of the glass in said tubular member, arising as a result of rapid temperature variations in said rod, over an appreciable length of wall of said tubular member, whereby excessive temperature gradicuts in said tubular member are prevented.
"I. A glass to metal seal comprising a rod of metal, a ring of glass completely surrounding said rod and adhering thereto throughout the circumference of said rod, the thickness of said ring at every point thereon being of an order of magnitude of less than about two millimeters, the upper edges of said ring being extended to form-a tubular glass body surrounding said rod but spaced therefrom, the thickness of said tubular body at the point where its spacing from said rod starts also being of said order of magnitude.
8. A glass to metal seal comprising a solid rod of tungsten having a diameter of the order of five millimeters or more, a ring of borosilicate glass completely surrounding said rod and adhering thereto throughout the circumference of said rod, the thickness of said ring at every point thereon being of an order of magnitude of less than about two millimeters, the upper edges of said ring being extended to form a tubular glass body surrounding said rod but spaced therefrom, the thickness of said tubular body at the point where its spacing from said rod starts also being of said order of magnitude.
9. A glass to metal seal comprising a metal body, a layer of glass adhering to the surface of said metal body, the upper edges of said layer being extended to form a glass body spaced from said metal body, said glass layer at every point thereon being not greater than where l=th'ermal conductivity for a unit cube of the glass in said layer,
P =tensile strength of said glass,
Ea=coefiicient of expansion of the metal of said rod,
En=coeflicient of expansion of said glass,
To=normal temperature of seal at which substantially no strains occur in said glass ring,
TA=maximum steady state temperature which said seal reaches during operation,
MB=Young's modulus for said glass,
S=thermal capacity per unit of mass of said glass, =density of said glass.
K=rate of temperature change of said metal rod of the order of magnitude of 20 C. per second or more.
the angle between the walls of said glass body and the walls of said metal body at the point where the spacing between said walls starts being sufficiently acute to prevent an increase in the thickness of the glass at said point as measured in a direction at right angles to the surface of said metal body.
10. A glass to metal seal comprising a tubular glass member closed at one end, a metallic rod sealed directly to the glass through said closed end, a powdered metal packed in the inside of said tubular member in contact with said seal, said conductor and the inner side walls of said tubular member, and means for retaining said powdered metal in intimate contact with all of said elements within said tubular member.
JAMES D. LE VAN.