Removal of long-lived 222 Rn daughters by electropolishing thin layers of stainless steel
R. W. Schnee, M. A. Bowles, R. Bunker, K. McCabe, J. White, P. Cushman, M. Pepin, V. E. Guiseppe
RRemoval of long-lived
Rn daughters by electropolishingthin layers of stainless steel
R.W. Schnee ∗ , M.A. Bowles ∗ , R. Bunker ∗ , K. McCabe ∗ , J. White ∗ , P. Cushman † ,M. Pepin † and V.E. Guiseppe ∗∗ ∗ Department of Physics, Syracuse University, Syracuse, NY 13244 † School of Physics & Astronomy, University of Minnesota, Minneapolis, MN 55455, USA ∗∗ University of South Dakota, Vermillion, South Dakota 57069, USA
Abstract.
Long-lived alpha and beta emitters in the
Rn decay chain on detector surfaces may be the limiting backgroundin many experiments attempting to detect dark matter or neutrinoless double beta decay. Removal of tens of microns ofmaterial via electropolishing has been shown to be effective at removing radon daughters implanted into material surfaces.Some applications, however, require the removal of uniform and significantly smaller thicknesses. Here, we demonstrate thatelectropolishing < µ m from stainless-steel plates reduces the contamination efficiently, by a factor > Keywords:
Radon daughters plate-out, electropolishing of steel surfaces, radiopurity
PACS:
INTRODUCTION
A particularly dangerous contamination for a number of rare-event searches or screening detectors is the depositionof radon daughters from the atmosphere onto detector components made of relatively clean materials such as stainlesssteel. These radon daughters decay to
Pb, a low-energy beta emitter with a long, 22-year half-life, and then to
Biand the alpha-emitting
Po. The
Pb daughter is usually plated onto surfaces or, due to the recoil energy receivedfrom
Po (and possibly
Po) decay, is implanted into a sub-surface layer of the material in question ( (cid:46)
50 nm forstainless steel).Electropolishing of stainless steel has been shown to be very effective at removing both
Pb and
Po [1].However, such electropolishing has removed significantly more material ( ∼ µ m) than is allowable for manyapplications. In particular, removing such a large thickness would be unacceptable for removing contaminationfrom the 25 µ m-diameter (125 µ m-diameter) stainless-steel wires constituting the anode (cathode) planes of theBetaCage [2, 3, 4, 5, 6]. In this paper we present the test results of electropolishing 30–1200 nm from the surfacesof stainless-steel samples that were artificially contaminated by exposing them to a strong radon source. PREPARATION AND ELECTROPOLISHING OF SAMPLES
Four unpolished (mill finish) 316 stainless-steel samples, 2 in. × × . × Bq m − day − . Measurement in an atomic-force microscope indicated a mean surfaceroughness of about 8.6 µ m. An additional, smoother sample ( ∼ µ m surface roughness), 1.5 in. × a r X i v : . [ phy s i c s . i n s - d e t ] A p r C Power Supply + – – Plas2c Rods Stainless-‐Steel Vessel Level of Acid Solu2on (40% H PO SO ) Copper Electrode (Cathode) Stainless-‐Steel Sample (Anode) A V
FIGURE 1.
Left : Schematic of the electropolishing setup showing the stainless-steel container and electrode-sample geometry.
Right : Photo of electropolishing setup within fume hood. surface roughness. After exposure, the exposed surfaces were not cleaned and all handling was done with glovedhands touching only the sides or bottom of each sample.The electropolishing was performed using the simple setup shown in Fig. 1. It consisted of a voltage source, currentand voltage meters, and two 100 mm ×
100 mm square cathodes made out of copper. The applied electrolyte wasa mixture of H PO (40%) and H SO (40%). The applied voltage was 2.4 V (DC). Although the current duringelectropolishing tends to decrease over time, the value was effectively constant ( ∼ ∼ ρ = . − , wereabout 4 nm/s for the square samples. MEASUREMENT OF UNIFORMITY AND REMOVAL OF CONTAMINATION
An alpha spectrometer (ORTEC Alpha Ultra-AS 33.8 mm diameter Si detector) with a background of 1 . ± . Po surface activity. A small distance between thedetector and the sample (5 mm) and low operating pressure of 200 mTorr allow for alpha spectroscopy with a highefficiency (13.8%) and good energy resolution, so long as the contamination is on the sample surface (see Fig. 2). Theabsolute efficiency was taken as the geometrical efficiency of the detector, and checked by comparing the measuredrate for a calibrated source to the expected rate based on the calculated geometrical efficiency. However, since finalresults depend only on relative changes in the signal from before to after cleaning, the uncertainty on the absolutecalibration of the detector is ignored here.Figure 2 shows the results of the electropolishing on the measured
Po alpha rates, ignoring (small) correctionsfor drifts in the digital scale, drifts in the alpha identification efficiency due to gain drifts, and for grow-in of the
Po over the time to make these measurements. The reduction approximately follows an exponential profile withcharacteristic mean thickness t ≈
50 nm causing a 1 / e reduction, with similar results for all samples with total ratesabove 10 alphas/day. For Sample Po is removed per 50 nm of material removed is the samplesurface roughness. Electropolishing tends to smooth out surfaces by preferentially removing atoms from protrusions,so much of the removed material is likely from parts of the sample deeper than any
Po. The first data fromthe smoother sample suggests contamination was removed at a slightly faster rate. Studies in the near future willconcentrate on determining the relationship between t and surface roughness by electropolishing smoother samples
000 5200 540010 (cid:239) (cid:239) Energy (keV) C oun t s / ( k e V d a y ) Mean Thickness Removed (nm) A l ph a s / d a y FIGURE 2.
Left : Raw spectra of Sample > × is apparent for events in the 5.2–5.4 MeV region of interest for Po. The rate after removal is consistent withthe background rate of the detector (when empty).
Right : Alpha detection rates for Sample × ’s), as a function of the mean cumulative thicknessof material removed. Uncertainties are smaller than the symbols if no error bars are shown. Lack of reduction in alpha rate from200 nm to 500 nm removed from Sample Po contamination in the rinse water. After freshultra-pure water was used for subsequent electropolishing, the alpha rate dropped to the background rate of the detector (yellowhorizontal band). Preliminary results shown here do not include (small) corrections for drifts in the digital scale, drifts in the alphaidentification efficiency due to gain drifts, nor for grow-in of the
Po over the time to make these measurements.
42 44 46 48 50 520100200300400500600 Energy (keV) C oun t s bkgsamplebkg fitsample fit
42 44 46 48 50 520100200300400500600 Energy (keV) C oun t s bkgsamplebkg fitsample fit FIGURE 3.
Left : Inside of GOPHER shielding, showing HPGe detector with sample directly on top of the Al window, surroundedby the Cu liner.
Center : GOPHER spectrum (red histogram) of Sample
Pb activity of 14 . ± . . Combined withalpha counting before electropolishing, this measurement indicates that the Pb activity for Sample . ± . beforeelectropolishing. Right : GOPHER spectrum (red histogram) of Sample
Pb activity of < . after electropolishing 1.2 µ m off. while correcting for systematic effects.Removal of Pb by electropolishing small thicknesses of steel was confirmed by observing the 46.6 keV gammaline with two samples placed directly on top of the 1.6 mm aluminum window of the GOPHER n-type high puritygermanium detector, as shown in Fig. 3. GOPHER sits within a high-purity copper-lined lead shield and includes aradon purge unit with sample load-lock. A detailed, custom GEANT4 simulation of the detector sample geometryindicates a 35% efficiency for 46.6 keV gammas originating on the surface of the sample facing the detector window.Samples µ m removed) and . ± . Pb. This measurement, combined with the relative alpharates of Samples
Pb activity on Sample . ± . . Measurement of Sample . The reduction factordue to electropolishing was ≥
12, consistent with expectations that the reduction of
Pb would be at least as large asthe measured factor of ∼ × for Po for the same Sample
IGURE 4.
Image of an originally 25-micron-diameter stainless steel wire before (left) and after (right) 60 seconds of electropol-ishing in the same setup shown in Fig. 1, which reduced the diameter about 2 µ m. Uniformity of electropolished wires
To test the uniformity resulting from electropolishing thin layers of stainless steel, a strand of 25-micron finewire (California Fine Wire Co.) approximately two inches in length was electropolished in the same setup shownin Fig. 1 for sixty seconds, cleaned with ultra-pure water and isopropyl alcohol, and dried with compressed CO gas. The wire was then measured at multiple points along its length using a scanning electron microscope andthe GIMP image manipulation program (see Fig. 4). An unelectropolished wire was examined and analyzed forthe control. The unelectropolished wire showed a standard deviation σ =
440 nm around its mean diameter µ = . µ m. The electropolished wire had σ =
600 nm and µ = . µ m. The relatively large standard deviation onthe unelectropolished wire may indicate that handling causing the wire to deform into an oval cross-section may bethe dominant source of measured non-uniformity. In any case, even for this relatively large thickness removed, theresulting uniformity is quite good. In particular, the standard deviation σ =
600 nm would be sufficient to provide <
10% gain variation in a drift chamber, sufficient for the proposed BetaCage [2, 3, 4, 5, 6].
ACKNOWLEDGMENTS
This work was supported in part by the National Science Foundation (Grants No. PHY-0855525, PHY-0919278) andthe U.S. Department of Energy (through Award Number DE-SCOO05054).
REFERENCES
1. G. Zuzel, and M. Wójcik,
Nuclear Instruments and Methods in Physics Research A , 140–148 (2012), URL .2. T. Shutt, C. E. Dahl, L. de Viveiros, R. J. Gaitskell, and R. W. Schnee, “Beta Cage: A New, Large-Area Multi-Wire ScreeningDetector For Surface Beta Contamination,” in
Topical Workshop on Low Radioactivity Techniques: LRT 2004 , AmericanInstitute of Physics Conference Proceedings, Melville, NY, 2005, vol. 785, pp. 79–83.3. R. W. Schnee, Z. Ahmed, S. R. Golwala, D. R. Grant, and K. Poinar, “Screening Surface Contamination with BetaCage,”in
Topical Workshop on Low Radioactivity Techniques: LRT 2006 , American Institute of Physics Conference Proceedings,Melville, NY, 2007, vol. 897, pp. 20–25.4. Z. Ahmed, S. R. Golwala, D. R. Grant, M. Kos, R. H. Nelson, R. W. Schnee, and B. Wang, “Status of BetaCage: anUltra-sensitive Screener for Surface Contamination,” in
American Institute of Physics Conference Series , edited by R. Ford,2011, vol. 1338 of
American Institute of Physics Conference Series , pp. 88–92.5. Z. Ahmed, M. A. Bowles, R. Bunker, S. R. Golwala, D. R. Grant, M. Kos, R. H. Nelson, A. Rider, R. W. Schnee, D. Sotolongo,B. Wang, and A. Zahn,
JINST , P01009 (2014).6. R. Bunker, M. A. Bowles, M. Kos, R. W. Schnee, B. Wang, Z. Ahmed, S. R. Golwala, R. H. Nelson, A. Rider, A. Zahn, andD. R. Grant, “The BetaCage, an Ultra-sensitive Screener for Surface Contamination,” in Topical Workshop on Low RadioactivityTechniques: LRT 2013 , edited by L. Miramonti, and L. Pandola, 2013, vol. 1549 of