US 20060001569A1 (19) United States (12) Patent Application Publication (10) Pub. N0.: US 2006/0001569 A1 Scandurra (43) Pub. Date: Jan. 5, 2006

(54) RADIOMETRIC PROPULSION SYSTEM Publication Classi?cation (76) Inventor: Marco Scandurra, Cambridge, MA (51) 110.0. 001s 3/02 (2006.01) 601K 15/00 (2006.01) Correspondence Address: (52) Us. 01...... 342/351;374/1 STANLEY H. KREMEN 4 LENAPE LANE (57) ABSTRACT EAST BRUNSWICK, NJ 08816 (US) Ahighly ef?cient propulsion system for ?ying, ?oating, and Appl. No.: ground vehicles that operates in an atmosphere under stan (21) 11/068,470 dard conditions according to radiornetric principles. The Filed: Feb. 22, 2005 propulsion system comprises specially fabricated plates that (22) exhibit large linear thrust forces upon application of a Related US. Application Data gradient across edge surfaces. Several ernbodi rnents are presented. The propulsion system has no moving (60) Provisional application No. 60/521,774, ?led on Jul. parts, does not use Working ?uids, and does not burn 1, 2004. hydrocarbon fuels.

1O 13

12 \ 11 Patent Application Publication Jan. 5, 2006 Sheet 1 0f 16 US 2006/0001569 A1

FIG. 2 Patent Application Publication Jan. 5, 2006 Sheet 2 0f 16 US 2006/0001569 A1

25 2X -

FIG. 6 Patent Application Publication Jan. 5, 2006 Sheet 3 0f 16 US 2006/0001569 A1

27 0000000. I.N... 000000.

FIG. 7 FIG. 8

FIG. 9 ~ FIG. 10 34

39 QI 40/? 9:

38 35 FIG. 11 Patent Application Publication Jan. 5, 2006 Sheet 4 0f 16 US 2006/0001569 A1

42

43/

FIG. 12 Patent Application Publication Jan. 5, 2006 Sheet 5 0f 16 US 2006/0001569 A1

52 6 6 7 a v 74 55 75 59 -l 4/ 155 56x 1 6 ‘ 70\IX - K.— —- | 0000000000000 0000000000000 ' 0000000000000 0000000000000 J 0030000000000 4J3 OOOOOOOOQOOOO 0000000000000 0000000000000 0000000000000 6 2,4 \ 0000000000000 \ x ’1_ ooooooooooo 00 '# H 60 ‘ 53 ' 1 ‘0000000000000000 00000000 ‘ 1 54 f I ‘ ‘0003000000001 r I 4 6 6 . x 5 62 6 |+ \6 1 asi a 2/ ,I FIE. 13 FIG. 14 76 M78

FIG. ,15 FIG. 16

/79 2 K5‘ I 8? {'86 I 89

8O K87

IF' ' K8100?“ " ‘\90 Wilt“91 5 FIG. 17 FIG. 18 Patent Application Publication Jan. 5, 2006 Sheet 6 0f 16 US 2006/0001569 A1

92

FIG. 19 Patent Application Publication Jan. 5, 2006 Sheet 7 0f 16 US 2006/0001569 A1 K106 108\ 102 \[———d—

K97 0000000000000 I l l l l 0000000000000 I I #11“ 0000000000000 4 0000000000000 0000000000000 0000000000000 0000000000000 0000000000000 0000000000000 0000000000000

0000 00000000 ~———___—— ---—-----4 |00000 0000000 \...__...,._....__J 110

FIG. 21 Patent Application Publication Jan. 5, 2006 Sheet 8 0f 16 US 2006/0001569 A1

111 117

11-1mm“I 00000000000 00000000000

115

1 1.4 00000000000 00000000000 "0.... 0000000000

111 120

121

123 124

0000066060 00000000000 0 00°00 0000000000 00000000000 00000000000 00000000000 0000000000 00000000000 00000000000 0000000000 00000000000 00000000000 0000000000 00000000000 00000000000 0000000000 00000000000 0000000000 00000000000 0000000000 00000000000 00000000000 l 00000000000 00000000000 ' 00000000000 00000000000 00000000000 125 00000000000 00000000000 122 00000000000 00000000000 °°°°°°§O°Q° 00000000000 00000000000 I‘ 00000000000 000000 00°00 I 00000000000 00000000000 |. 0000000000 00000000000 00000000000 0000000000 00000000000 00000000000 F vuvvvvvvvv 90000909990 00000000000 n 0000000000 00000000000 00000000000 I 0000000000 00000000000 0000000000 00000000000 00000000000 I. 0000000000 00000000000 00000000000 00000000000 0000000000 00000000000 00000000000 0000000000 00000000000 0000000000 00000000000 00000000000 0000000000 00000000000 0000000000 00000000000 0000000000 00000000000 0000000000 00000000000 0000000000 0000000000 0000000000 0000000000 00000000000 00000000000 0000000000 00000000000 00000000000 FIG. 24 Patent Application Publication Jan. 5, 2006 Sheet 9 0f 16 US 2006/0001569 A1 K 129 K 127

‘X6 \ 130 128 FIG. 25 Patent Application Publication Jan. 5, 2006 Sheet 10 0f 16 US 2006/0001569 A1

139 141 /

4.4

1.- f5it_: “Ma~.__,"“3 JJK4Jim4J5.4:Li5 Z a.+

K -----I 1 3 OOOOOODPOOOOOOODOOODOOOOOOOOOOOOOOO O00000OOOOOOOOOOOOOOOOOOOOO00000000 OOOOOOOOOOOOOOOOOOOOOOOO00000000000 OOOOOOOOOOOOOOOOOOOOOOOOO0000000000 OOOOOOOOOOOOOOOOOOOOOOOO000000000O0 0006000000000OOOGOOOOOQOOOQOOOOOOOO 00000000000OOOOOOOOOOOOOOOOOOOOOOOO 0O0O00000000OOOOOOOOOOOOOOOOOOO0000 00000000OGQOQQOOOOOGOOOO00000000000 OOOPDDQOQGOOOQDDOOODOQOOODODOOOOOO OOOOOOOOOOOOOOOOOQOOOOOOOOOOOO00000 0000OOOOOQOOOOOOOOOOOOOOOOO000000O 0OOOOOOOOOOOOOOOOOOOOOOOO00000000O OOOOQOOQOOOOOOOOOOOOOOOO000006000 0000000000OOOOOOOOOOOOOOOOOOOOOOOO 000000000000000OOOOQOOOOOOOOOOOOOOO 000000OQOOOOOOOOOQOOOOOOOOO0.0000000 OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO ‘ 00000000000OOOIODObOOOOOODOODOOOOO OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO OOOOOOOOOOOOOQOOOQOOOOOOOOOOOOOOOOO OOOOOOOOOOOOOOOOOOQOOOOO00000000000 OOOOOOOOOOOOOOOOOOOOOOOO00000000000 OOOOOOQOOOOOOQOOOOOGOOOO00000000000 OOOOOOOOOOOOQOOOOOOOOOOO00000600000 OOOOOOOQOOOOOOOOOOOOOOOO00000000000 OOOOOOOOOOOOOOOOOOOOOOOO00000000000

r 134 FIG. 27 [[141 i

k If \142 137 136 135

138 FIG. 28 Patent Application Publication Jan. 5, 2006 Sheet 11 0f 16 US 2006/0001569 A1

143

4| 4 4

o-. oro-00-LoOh_F www“aw?_www_0_. o50b00-\-InP ll\ 0N‘0001000\|u|J04000\ +.000.r__++_000_+_00000;000. C600_.000_+009+.000.0 4| 4 5 .04.0|_,\000IN."03.01. 000000w:6.000..Lw000C +_000_000 0“4000“#0400? LO-.00CWC.- - OWH0OO\JOOOi#04OOO\OOOQ 702_000_000_o.0 11100911100000/114x0001..\000 @000‘ _00o__0o0_0300-0. ooo?ntoorwltooof:000.1“;000.1“._ 000/101.1.210001.11.QOOIJIJXOOOI]0c0c\ + . 4000_M0007000“JO JOOO\ OQQOQOQOOOOOQOOOOOOOOOOOOOQQOOQQQO 0 OOOOQOQOOOOOQOOOOOOOOOOO00000000990 00°0°OaoooooooonvoonuooQOOOOQQOQOOOQo OQOOOQOOOQQOQCOOQQQOOOOOonvoooQOonvowl OOOQOQQOQOOOQOOOOOOQOQQOOOQQOQOQOQC O0°0°O0O000000000OOOQOQOOOOOOOQOOOO 00000000000000000000000000000000000 00°QOOQOQQQOQOQQOQQOCOOCOOGOOOQOQOQ 00°0°Q0O000000000OQOOOQQOOOCOOOOQOO 00°000°C0O9000000OOOOQOOOOOOOQOQOOO QQOQQOQOOOQOOQQQQQQQOOQOQOOQOOOOOOO 00°oocoaoooocoonvonvoooccoQCOQCQQCQQO 00000000000000000000000000000000000 OOOOOOOOOOOOOOOOOOOOOOOO00000000000 00000000000000000000000000000000000 00000000000000000000000000000000000 00000000000000000000000000000000000 0000000000O000000000000000000000000 00000000000000000000000000000000000 00000000000000000000000000000000000 00000000000000000000000000000000000 000000000000000000000000000000000004 00000000000000000000000000000000000 00000000000000000000000000000000000 00000000000000000000000000000000000 00000000000000000000000000000000000 00000000000000000000000000000000000 FIG. 29

146 143

i

148 149 FIG. 30 Patent Application Publication Jan. 5, 2006 Sheet 12 0f 16 US 2006/0001569 A1

157 \ 155

152

153 154 150

156

FIG. 31 1:;9 /-160 /158 g 1?

155

K16" FIG. 32 159

Patent Application Publication Jan. 5, 2006 Sheet 14 0f 16 US 2006/0001569 A1

169 171 J 170. [173

2/ ' 170‘ 17

174

FIG. 35 Patent Application Publication Jan. 5, 2006 Sheet 15 0f 16 US 2006/0001569 A1

188 \‘i 7/175 f179 7/ 187 186 15 7/ ' 7 K, 7; , / 17 189%-\ 182-\ / 183 &\\// t1 \176 180

FIG. 36 Patent Application Publication Jan. 5, 2006 Sheet 16 0f 16 US 2006/0001569 A1

llliltlnmb. fllll illl‘ljllllllllilillllii I‘lHllill.' Ill-Ill if!‘

j 21 US 2006/0001569 A1 Jan. 5, 2006

RADIOMETRIC PROPULSION SYSTEM generates a force directed toWard the colder silver surface as air molecules contained in the vessel impinge on the vanes. CROSS REFERENCE TO RELATED In fact, air molecules at loW density exert different pressures APPLICATIONS on hot and on cold bodies. HoWever, the force driving the radiometer is small, of the order of 10-6 N, and furthermore, [0001] This nonprovisional patent application is a at atmospheric pressure the effect vanishes. The present National Stage Entry under 35 U.S.C. § 371 of International invention addresses and solves these disadvantages. Abrief PCT Application Number PCT/US 05/02820 ?led on Jan. 31, 2005. This application claims the bene?t of and priority explanation of the driving mechanism of the Radiometer is to this PCT application Which, in turn, claims the bene?t of necessary in order to understand the method of action of the and priority to US. Provisional Applications Nos. 60/481, present invention. Such an explanation can be found in more 999 ?led on Feb. 2, 2004 and 60/521,774 ?led on Jul. 1, detail in the textbook: “Kinetic Theory of gases”, by L. B. 2004 by the Present Inventor. The Present Application also Loeb, 1961 edition, Chapter VII section 84. claims the bene?t of and priority to both Provisional Appli cations. The PCT Application and both Provisional Appli [0007] At equilibrium air can be considered to be a gas cations are incorporated it herein in their entirety hereto. consisting of molecules that are freely moving, point-like particles. Each particle moves With a mean velocity V BACKGROUND proportional to the square root of the temperature. Thus molecules move faster and maintain a higher [0002] 1. Field of the Invention When the air temperature is higher. Molecules bouncing off a hot surface are repelled from that surface at a higher speed [0003] The ?eld of the invention is concerned With pro pulsion systems, rare?ed gas dynamics, and nanotechnol than molecules bouncing off a cold surface. Therefore, one Would expect molecules to impart a higher momentum to hot ogy. surfaces than to cold surfaces. Consequently, the forces on [0004] 2. Analysis of the Problem and the Prior Art hot surfaces are greater than on cold surfaces. As shoWn in [0005] Propulsion systems for modern aircraft, boats, FIG. 2, in a device having tWo surfaces 6 and 7, separated ships, submarines, and automobiles are based on macro by an insulator 8, said surfaces being at tWo different scopic rotary motion of the engine components such as : Surface 6 is at temperature To (cold); Surface propeller, turbine blades and Wheels. These systems, due to 7 is at temperature Th (hot); Where Th>TC said device being the presence of massive gears, shafts and other moving surrounded by air at temperature T, one should observe a parts, introduce poWer losses due to , slippage, and force that pushes the device toWard the cold surface. One heat; and reliability problems and maintenance costs related Would expect this force to be proportional to the area of the to mechanical Wear. Airplane turbines expel hot gases hot and cold surfaces and to the number of particles present through a noZZle, thereby limiting their use in the global in a unit volume of air. This force Would also be expected to environment for safety reasons. Large moving propeller be larger at atmospheric pressure than in a partially evacu blades also represent a haZard to people approaching an ated vessel. These expectations are contrary to experimental operating engine. Another disadvantage of a propulsion evidence. A thin glass WindoW that is cold on the outdoor system based upon macroscopic motion and combustion is surface and Warm on the indoor surface Will not bend nor the considerable noise and air pollution. Large sums of break due to this force. The error in the model is in the money are being expended in annual research by private and assumption of freely moving, point-like particles. In the governmental agencies in the effort to reduce acoustic presence of temperature differences and in dense air, colli pollution due to airplane turbines. Noise pollution also sions betWeen the gas molecules cannot be neglected. After forces airlines to spend considerable sums of money to striking the cold surface, a molecule cannot depart quickly reimburse urban communities adjacent to airports. This from that surface because it is trapped by the surrounding translates into increased airfares. Current air propulsion molecules. Therefore, the air above a cold surface is cold, systems also require high poWer thereby necessitating the and the air above a hot surface is hot. Furthermore above a use of fossil fuels for hydrocarbon combustion thereby cold surface the gas is denser, and the number of molecules contributing to environmentally dangerous gas emission and striking the cold surface is larger thereby enhancing the global Warming. It Would be desirable to have a propulsion momentum imparted to the cold surface. Above the hot system having no moving parts, no operational noise, and surface the gas is less dense, and the number of molecules that Would be ef?ciently poWered by electric batteries or fuel striking this surface is smaller, thereby reducing the momen cells. tum imparted to the hot surface. This effect compensates for the radiometric force. In fact, pressure in the gas equaliZes [0006] Such an alternative is suggested by a fragile engine throughout, thereby preventing any signi?cant force from called a Crookes Radiometer. FIG. 1 shoWs a representation arising outside the Crookes vessel. of a Crookes Radiometer. This device is comprised of a partially evacuated bulb 1, a pivot 2, and a four Winged mill [0008] If molecules at room temperature could reach the 3 mounted on pivot 2. Each Wing or vane is lamp-blacked on surface and leave it quickly after collision, the radiometric one side 4, and silvered on the other side 5. When intense mechanism described above could Work in air at standard impinges on the Wings the mill starts spinning due to temperature and pressure. This mechanism hoWever, is only a radiometric force. The motion is completely silent. The realiZed When the air is rare?ed, and in particular When the method of action of the radiometric force is roughly molecular mean free path is larger than the siZe of the described as folloWs. The black surface 4 of each vane surface. The mean free path 7» is the average distance that a becomes hotter than the silvered one 5 because of the molecule travels before encountering another molecule. It is different absorption coef?cients. The temperature difference given by: US 2006/0001569 A1 Jan. 5, 2006

ef?cient at edges and tips. O. Reynolds and G. Hettner recogniZed the existence of the How of molecules from the 1 [1] cold to the hot side of the vane (Phil. Trans. Roy. Soc. Lon., 166, 725, 1876, and Z. Physik, 27, 12, 1924). The ?oW velocity is linearly dependent on the temperature gradient 6T/6x. The How occurs at the edge of the vane Where a Where: (I is the diameter of the molecule; and, strong temperature gradient exists. Associated With this streaming of molecules is a force directed opposite to the [0009] n is the number of molecules present in a unit temperature gradient. Consider the vane shoWn partially in volume. FIG. 4. The vane 15 extends in the y-Z (facial surface) plane, [0010] Since 7» is inversely proportional to the air density, the thickness L extends along the x (edge surface) direction. the molecules move almost Without collisions at suf?ciently The cold facial surface 16 is kept at loW temperature To and small values of n. At these densities, the model of freely the hot facial surface 17 is kept at high temperature Th such moving, point-like particles applies, and the radiometric that (Th—TC)=AT. The creep force 19 pushes edge surface 18 force is governed by the mechanism described above. A in the direction shoWn, such force is opposite to the How 20. device exploiting this force is Knudsen’s Absolute Manom A formula for this force can be found in section 82 of the eter, a pressure gauge described in section 82 of the book: book: “The Kinetic Theory of Gases” by L. B. Loeb (Dover, “The Kinetic Theory of Gases” by L. B. Loeb (Dover, NY, NY, 1961) as quoted from the article by Hettner et al. and 1961). The absolute manometer described therein utiliZes it is given beloW in terms of the parameters relevant for the the radiometric effect to measure gas pressure in a high present invention, namely the molecule average diameter a chamber. Here, the force is proportional to the the molecule density n and the thickness of the vane L: surface area. [0011] The Crookes Radiometer of FIG. 1 also utiliZes a rare?ed gas environment. HoWever, the pressure inside Crookes vessel is typically of the order of 0.1 mm Hg. Under these conditions the molecular mean free path is about 0.1 mm, While the side b (the linear dimension) of each vane is larger than 1 cm. Therefore, the radiometer Works at rela Where k is the BoltZmann constant and SV is the area of tively high gas densities for Which the free particle model surface 18 shoWn in FIG. 4. The creep force is proportional does not apply. The theory of radiometric forces at these to the surface of the vane Which is parallel to the temperature densities Was elaborated in the course of intense theoretical gradient. and experimental investigations by many authors in a period [0014] Thermal creep, also called thermal transpiration, stretching from 1874 to 1926. It Was discovered that the Was ?rst observed by O. Reynolds (Phil. Trans. Roy. Soc. force is in part a shear effect due to thermal creep of air from Lon. 1876), and Was used by him to generate a pressure the cold to the hot surface of the vanes, and is in part a difference betWeen tWo chambers containing gases kept at “facial effect” due to a local pressure difference as discov different temperatures. The chambers Were separated by a ered by (“Zur Theorie der Radiom thick porous plug made of stucco or Meerschaum. Such a eterkrafte,” Berlin, July 1924). plug or membrane comprises long channels having a very [0012] Radiometric forces arise in a dense gas When a small diameter. The gas contained in the cold chamber temperature gradient is present, they are proportional to the moved by creeping along the channels toWard the hot intensity of the gradient and they are alWays directed oppo chamber. The How Was very sloW, but Reynolds achieved a site to it, i.e., toWard the cold region. The forces act on small measurable pressure differential betWeen the tWo chambers solid bodies—a mean free path Wide—immersed in the gas. When the temperature difference Was of the order of one When a large body is immersed in the gas, only a limited hundred degrees Fahrenheit. This effect Was exploited by portion (of siZe 7») of the body is radiometrically active. The several inventors to fabricate vacuum pumps. One of the unequally heated plates inside Crookes Radiometer provide latest is a micro pump patented under US. Pat. No. 6,533, a simple example. They generate a temperature gradient in 554. This device utiliZes a silica aerogel membrane as a the air contained in the vessel. The gradient is stronger at the porous plug. Aerogel membranes have a channel-like struc edge of the plate and therefore the radiometric force is ture With a channel diameter smaller than the mean free path stronger here, as shoWn in FIG. 3. Here, a vessel 9 encloses of air at atmospheric pressure. The aerogel also has a loW a radiometric vane 10, mounted on a pivot 11. The vane is thermal conductivity Which helps to insulate the tWo cham hot on the loWer surface 12 and cold on the upper surface 13. bers. This device produces a loW gaseous ?oW With the Above the central regions of the vane there is a small remarkable advantage of having no moving parts. Recently temperature gradient. HoWever, at the vane edge there is a the use of a thermal transpiration pump as a part of a large gradient 14 due to the proximity of the hot and cold micro-scale propulsion system Was proposed by F. Ochoa, surfaces and this is the active region. This region is enclosed C. EastWood, P. D. Ronney, and B. Dunn (“Thermal Tran by a phantom line in FIG. 3 spiration Based Microscale Propulsion and PoWer Genera tion Devices,” 7th International Microgravity Combustion [0013] J. C. MaxWell ?rst pointed out in 1866 that the Workshop, Cleveland, Ohio, June 2003). They disclose a radiometric force acts only at the edge of the vane, and that thermal transpiration pump very similar to the one used in most of the vane surface is inactive (Phil. Trans. Roy. Soc., the above mentioned patent. The pump is utiliZed to com pp 49-88, London 1866). According to MaxWell, the radio press reactants before combustion in a combustion chamber metric forces derive from heat exchange, and are thus more in order to obtain thrust. This microscale propulsion system US 2006/0001569 A1 Jan. 5, 2006

has the advantage of no moving parts. However its disad [0021] Einstein calculates the force for the case of a plate vantages include the following: Which is hot on one side and cold on the other (as in Crookes [0015] 1. Use of combustion of hydrocarbons to drive radiometer) by inserting into equation [3] the folloWing the compressor, as in conventional turbine propulsion, gradient With consequent production of hot haZardous gases to be ejected by a noZZle. [0016] 2. Emission of pollutant gases. [0017] 3. The large thickness of the membrane prevents generation of large creep forces. Aerogel sheets beloW 0.1 mm thickness are dif?cult to fabricate and Would be [0022] In reality, Einstein’s model uses a plate With no too fragile to sustain stresses and pressure differentials. thickness, and thus the gradient Would be in?nite. HoWever Einstein assumes that 7» is the minimum effective thickness [0018] 4. Operational temperatures of several hundred and therefore, equation. [4] yields the maximum effective degrees Kelvin above room temperature are necessary gradient. The force does not increase for L<7». Furthermore, to obtain usable forces. the concept of temperature is meaningless beloW the mean [0019] 5. The device produces a very limited thrust and free-path dimension. As a consequence of this, Einstein ?nds features a small thrust/Weight ratio. a radiometric force: [0020] There is also another reason for the existence of radiometric forces. It Was discovered by A. Einstein in 1924 T — Tc 5 (previously cited; see also L. B. Loeb, section 84 Where FEinstein = —P/\hT( [ 1 Einstein’s original paper is almost entirely translated). Ein stein found that radiometric forces also act on surfaces normal to the temperature gradient, i.e., on facial surfaces, [0023] Equation [5] explicitly shoWs the linear depen 16 and 17, as shoWn in FIG. 4. Einstein’s forces act on a tiny dence of the force on the perimeter 1. This dependence Was portion of the facial surfaces, said portion being only a mean experimentally veri?ed by Marsh, et al. (J.O.S.A. & R.S.I., free path Wide. Einstein’s approach is to consider an in? 11, 257, 1925 and JOSA & RSI, 12, 135, 1926). In these nitely thin vane as shoWn in FIG. 5. Here, an ideal radiom experiments the radiometric force Was measured for tWo eter vane 21, With arm 22 is shoWn. According to Einstein, kinds of plates shoWn in FIG. 6. The ?rst, a regular plate, the tiny strip 23 delimited by the phantom line is the active similar to a Crookes vane, is shoWn in FIG. 6 as element 24 area of the vane. The Width of the strip is approximately 7». (top vieW). The second plate 25 has a larger perimeter With Einstein’s theory can be summariZed as folloWs. When a a surface area equal the surface area of the plate 24. The large in?nitely thin plate is immersed in a region of air Where plates Were illuminated on the lamp-blacked side and the a temperature gradient 6T/6x exists, the pressure is constant force Was measured by means of a torsion pendulum. It Was in the column of air above the center of the plate—i.e., the found that for the plate With the larger perimeter 25 the force quantity nT is a constant. This is empirically observed. AWay Was larger than that for the plate With the smaller perimeter from the plate, Where no object is present, one observes that 24. Einstein’s equation [5] Was qualitatively veri?ed With the air does not move. Therefore the mass motion of the air reasonable accuracy—the discrepancies being attributed to must be Zero, and the quantity \/T is a constant. This the dif?culty in measuring and maintaining the temperature quantity and the quantity nT cannot be mathematically gradient in the perforated plates. constant at a same time. Thus, as Einstein points out, far aWay from the plate there is a net momentum ?oW, i.e. [0024] The idea of perforating the plate as in FIG. 6, With molecules coming from the hot side carry more momentum the goal of increasing the perimeter and thus the force, Was than molecules coming from the cold side. The molecules exploited by inventor Marc S. Paller (US. Pat. No. 4,397, from this region impinge obliquely on the facial edge of the 150, Aug. 9, 1983). Paller discloses a lattice of strips used in plate exerting a net force. In other Words, at the edge of the a radiometric light-mill Within a partially evacuated vessel. plate, there exists a transition region (23 in FIG. 5) of the The perforated vane produces a higher rotational speed of siZe of the mean free path 7» in Which there is a pressure the mill. The goal of the invention is to convert radiant differential. The pressure is higher on the side of the plate energy, (necessary to maintain the temperature gradient) into facing the higher temperature. Astrip of Width 7» on the facial rotational kinetic energy, and then into electrical energy. The surface of the plate Will thus experience a higher pressure gas to be contained in the vessel is possibly helium due to from molecules impinging obliquely from the hot side of the its small molecular siZe and large mean free path. The gas. The net force is proportional to the perimeter of the disadvantage of this apparatus is the need for a vessel plate and can be easily calculated by kinetic theory of gases containing a controlled gas type and density. It is a device and is given by: that requires a selection of gas type and gas conditions in order to ?t the device speci?cation. It Would be desirable to have a radiometric device that adapts to the natural occur ring conditions of air at atmospheric pressure. It Would also be desirable to have a radiometric device that produces a linear thrust rather than a rotational force on a carefully balanced rotational assembly—a thrust that is large enough Where 1 is the perimeter of the plate and p is the average gas to provide propulsion for vehicles. Generation of linear pressure. This force acts on surfaces normal to the tempera thrust is not the goal of the above mentioned invention by ture gradient (facial surfaces). Paller.