Status of the Knudsen Compressor for Use in Distributed and Autonomous Sampling Systems

Transcription

1 Status of the Knudsen Compressor for Use in Distributed and Autonomous Sampling Systems M. Young, Y.- L. Han, E. P. Muntz, G. Shiflett University of Southern California A. Green, S. Jones, Jet Propulsion Laboratory 4 th Harsh-Environment Mass Spectrometry Workshop St. Petersburg Beach, FL Oct. 7-10, 2003

2 Overview Requirement for Micro-scale Gas Roughing Pumps Introduction to Thermal Transpiration The Knudsen Compressor Knudsen Compressor Performance Models Radiantly Driven Knudsen Compressors Knudsen Compressor Cascade Experiments Perforated Aerogel Transpiration Membrane Experiments Optimization Results Sample Compressor Sizing Summary

3 Micro/Meso-Scale Gas Roughing Pumps Micro/meso-scale gas sensors are being developed that require gas pumping. Both roughing pumps and high vacuum pumps are required. Two possible development strategies: shrink down existing technology or develop new technology. Problems with shrinking down existing technology: Moving parts, required manufacturing tolerances, oil. Worth considering new pump technology for micro/meso-scales. A pump based on thermal transpiration is one promising technology for micro/meso-scale gas roughing pumps: No moving parts, no oil or supplementary fluids, scalable, similar technology applicable to high-pressure gas compressors, variety of powering options including: radiative, solar, resistive, waste heat, combustion.

4 Thermal Transpiration (Thermal Effusion and Creep) Rarefied gas phenomena (free-molecular or transitional flow driven by surface temperature gradient) Thermal Effusion Through Orifice T 1 <T 2 Thermal Creep along surfaces P 1,T 1 P 2,T 2 p 1 = p 2 T T 1 2 Longitudinal Wall temperature gradient drives creep flow, counterbalanced by pressure driven return flow (Poiseuille flow) One of the driving mechanisms in Crooke s radiometer Net effect is; for P 2 = P 1 a Maximum flow from cold to hot, for P2 > P1 Reduced Flow Reaching Zero for Maximum P 2 /P 1

5 Using Thermal Transpiration Flow in a Knudsen Compressor is the difference bet ween thermal creep and pressure driven return flows P TMPD = T P T TMPD FM = ½ TMPD Cont = 0 Rarefied flow in the capillary section (i-1)th Stage Capillary radius Lr,i T Tavg TL,i Lx,i Capillary Section Continuum flow in the connector section ith Stage LX,i x Connector Section (i+1)th Stage Ti/2 Ti/2 LR,i P (pi-1)eff pi κi pmax,i pc,i (pi)eff

6 Knudsen Compressor Membrane (Aerogel) Models Aerogel Gas Flow and Thermal Transpiration Model Aerogel Thermal Model ρ ρ a t Important Parameters π = 6 A pores 3 L L Π = 1 r r d d a a = Π ρ ρ por S s (m 2 /g) t A 3

7 Qp C LT Knudsen Compressor Flow Model Flow properties are dominated by transpiration membrane properties Gas Conductance k T C = V & P avg P C ST 1 = C 1 b avg 2L r = A Qp = + ( 2 Kn C ) r a Cma Ci Cma m Lx Qp,Sone Qp,cf Qt,Sone Qt,cf + 1.E-02 1.E+00 1.E+02 1.E+04 1.E+06 Kn Qt C i = A k b C T m (Including short tube effects) ma avg = p A LT + 2γ γ 1 1 C a 1 ( γ ) X γ 1 γ p X p p k b T avg 2πm 1/ 2

8 Knudsen Compressor Membrane Thermal Model Radiantly heated transpiration membrane. Optical energy is absorbed throughout the body. I x = I o e ( k +σ ) ρx λx v v = I o e I L 1 T + k λl o λx λl ( x) = x L + { e e } Tc x = 0 x = L Energy is lost through conduction through the material, radiation outward and free convection outward. = k ( T T ) amb n φ fc C fc Ra ( ) h ambient 4 4 φ = εσ T T L Thermal energy conducted through the porous material is the sum of the radiation, solid conduction, and gas conduction. k = k + k + t s g k r, si rad k s = c si ρ αa por h k g c I x φ fc φ rad k g, c = 1 + 4βKn r 16σT k r = 3λ k t 3 rad

9 Knudsen Compressor Membrane Transpiration Model Calculate Pressure Difference and Throughput for Membrane T QT p = pavg κ m T Q T M = ( κ ) Tavg Q & Ctm Pavg 1 P ktavg Tavg QP 1E+17 Ndot (#/s) 9E+16 8E+16 7E+16 6E+16 5E+16 4E+16 3E+16 2E+16 1E+16 κ = 0 N dot,max = 7.80E16 #/s P max = Torr κ = P (Torr) Same Analysis Works for the Connector Section, but is Neglected in the Current Work

10 Single Stage Knudsen Compressor (MEMS) Radiant Heating Used to Simplify Manufacturing Transpiration Membranes Aligned to Simplify Heating Pyrex Cover Silicon Thermal Guard Aerogel (in red) (8mm x 10mm x 1mm) Pyrex Bonding Window Kovar Inlet/Outlet Parts Anodically Bonded Design Optimizations Can Be Tested With Conventionally Machined Version

11 Conventionally Machined Knudsen Compressors Much Cheaper to Fabricate Than MEMS Version Very Similar Geometry O-Ring Seals Allow Multiple Transpiration Membranes to be Tested in Same Device Single Stage Cascade of 5 Single Stages 15 Stage Cascade

12 P Sensor Experimental Cascade Setup Separating Valve P c Sensor Knudsen Compressor Cascade Pump filled with pure gas Light Source turned on Separating valve closed Pressure rise vs. time measured

13 Experimental Process and Analysis P (Torr) Pc (Torr) Ph (Torr) DP (Torr) P (Torr) time (s) Ndot (#/s) 1E+17 9E+16 8E+16 7E+16 6E+16 5E+16 4E+16 3E+16 2E+16 1E+16 0 Pump Down y = -1.53E+16x E+16 y = -1.50E+16x E+16 Run 1 Run 2 Linear (Run 1) Linear (Run 2) P max,1 = 5.86 P max,2 = 5.75 N dot_max,1 = 8.97E16 N dot_max,2 = 8.63E P (Torr) Ndot/n (m 3 /s) 4.0E E E E E E E E E+00 Vent Up y = 4.63E-07x E-10 y = 4.67E-07x E-10 C = 4.65E-7 m 3 /s Run 1 Run 2 Linear (Run 1) Linear (Run 2) P/P avg

14 Experimental Results Single Stage T h (Torr) Membrane Temperature P (Torr) He produces a smaller P due to a smaller T from increased k g and φ fc φ= 45 mw/cm 2 P (Torr) N Th (C) Th,an (C) DP-He (Torr) DP-N2 (Torr) DP-N2,covered (Torr) Consistent underprediction of T h due to overprediction of φ r 1 Stage Steady-State P φ = 45 mw/cm P h (Torr)

15 P (Torr) Experimental Results 5 Stage Cascade 5 Stage Steady-State P φ = 15 mw/cm P c (Torr) 1.4E+17 Difference in throughput also depends on gas conduction He has the highest throughput, but lowest P Ndot,max 1.2E E E E E E E+00 He N2 Ar Difference in P due primarily to differences in cooling and T 5 Stage Maximum Throughput φ = 15 mw/cm P c (Torr) He N2 Ar

16 Experimental Results Summary 1,2,5 Stages P (Torr) DPmax,1 DPmax,2 DPmax,5 Ndot,max,1 (#/s) Ndot,max,2 (#/s) Ndot,max,5 (#/s) P c (Torr) 1.E+17 9.E+16 8.E+16 7.E+16 6.E+16 5.E+16 4.E+16 3.E+16 2.E+16 1.E+16 0.E+00 For these conditions throughput is relatively constant for various numbers of stages (dependence is due to manufacturing differences between the different stages) For these conditions the pressure difference scales with the number of stages N dot (#/s)

17 Experimental Results 15 Stage Cascade Ndot (#/s) 6E+16 5E+16 4E+16 3E+16 2E+16 Steady Heating Transient Heating Linear (Steady Heating) Linear (Transient Heating) φ = 15mw/cm 2 1E P (Torr) Approximately the same throughput as for 1,2,5 stages P of 8 Torr per stage same as for 1,2,5 stages

18 Perforated Aerogel Low Pressures Optimal operation requires Kn ~ 1 Array (9x11) of 380 µm holes drilled through aerogel transpiration membrane. Silica Aerogel Transpiration Membrane (80mg/cc, 8% carbon) Torr Seal Epoxy Aluminum Thermal Guard P max Still Underpredicted by ~ 15-20%. Proof of Concept Shown. P (mtorr) φ = 121 mw/cm 2 L r = 380µm Testing Conventionally Drilled Holes to L r = 250µm. Testing Laser Drilled Holes for L r = 250 µm to 25 µm. Experimental Analytical Poly. (Analytical) Poly. (Experimental) P (Torr)

19 ℵ = Future Optimization Cascade Operation Minimize energy consumption per unit throughput and pressure ratio Q& P N& P avg 1.E+06 φl = T T 1 2L Q 8km Pavg 1 ( ) κ κ ℵ x p 2 2 r Qt avg 2 nd Bracketed Term (Normalized) 1.E+05 1.E+04 1.E+03 1.E+02 1.E+01 1.E mtorr 100 mtorr 1 Torr 1.E-01 1.E-03 1.E-01 1.E+01 1.E+03 1.E+05 1.E+07 Kn Conclusion: Operate with Kn ~ 1 to minimize energy consumption

20 1 st Bracketed Term Optimization Cascade Operation II f = 50mw /cm2 f = 100mw /cm2 f = 200mw /cm2 f = 500mw /cm2 f = 1000mw /cm L (mm) φl = T T 1 2L Q 8km Pavg Conclusion: Operate at L x ~ 0.5mm, φ = 1000 mw/cm 2 1 ( ) κ κ ℵ x p 2 2 r Qt avg Conclusion: Operate with k = 1/2 to minimize energy consumption 1/(κ-κ 2 ) κ

21 System Sizing Matched to Creare s Turbopump Design Meso-Scale Gas Pumping System Operating on Air from 1E-5 Torr to 1 atm. High Vacuum Size Power Flow Rate Backing Pressure Lifetime φ = 5cm 1 W 5 1E-5 Torr 200 mtorr 1 Year Continuous Operation Roughing Pump Based on Radiantly Driven Knudsen Compressor Design Based on Optimization Analysis Calculation Made Using Experimentally Validated Knudsen Compressor Performance Model P = 200mTorr to 1 atm. N & = 1.7E15 mol/sec (5 1E-5 Torr) φ = 1000mw/cm 2, L x = 0.5mm, ρ = 80mg/cc, 8% carbon doped Cascade Characteristics STAGES = 258 POWER = 300 mw VOLUME = few cm 3 Stated power does not include conversion and distribution inefficiencies.

22 Experimentally Demonstrated: Cascades of up to 15 stages. Current Status Cascades operating on N 2, He, Ar, Air from P = 250 mtorr to 1 atm. Radiant Driving at φ = 15mw/cm 2 to 125 mw/cm 2 (including direct solar illumination). Knudsen Compressor Performance Code Validated Over Experimental Conditions. Required MEMS meso-scale manufacturing processes shown.

23 Required Future Work Proof of Concept Demonstration Continue demonstration of perforated aerogel transpiration membranes. Demonstrate high flux (1000 mw/cm 2 ) operation. Manufacturing Processes Stage miniaturization (1cm 2 < 1mm 2 ) Aerogel bonding process Aerogel sealing process Packaging and feedthroughs Make practical comparisons of different candidate heating techniques

24 Summary Micro/meso-scale Gas Roughing Pumps Based on Thermal Transpiration Can Efficiently Operate to a Pressure of 10 mtorr. Knudsen Compressor Performance Model Has Been Built and Experimentally Validated and Used in Optimizing Knudsen Compressor Designs. Cascades of Up to 15 Conventionally Machined Stages Have Been Designed, Built, and Tested. Experimental Results for the Gas Throughput, Pressure Difference, and Temperature Difference Agree with Model Results to Within 15%. Optimized Knudsen Compressor Designs for a Gas Roughing Pump Appear Viable. Initial Experimental Results with Etched Aerogel Transpiration Membrane Agree with Expectations.