An Annotated Bibliography of the "Reifenschweiler Effect": Temperature & concentration dependence of tritium activity in a titanium matrix ------------------------------------------------------------ Last revised: 26 December 1994 C. Harrison harr@netcom.com ------------------------------------------------------------ This document is available at file://sunsite.unc.edu/pub/academic/physics/cold-fusion/TiBib.txt Table of Contents: Sec. 1........Reifenschweiler works Sec. 2........Titanium hydride phase diagram Sec. 3........Surface interactions of Ti, Hydrogen, and Oxygen Sec. 4........H diffusion in Ti: NMR studies, "hydrogen pairing" Sec. 5........Tritium-specific data ================================= Section 1. Reifenschweiler works ================================= [Reifenschweiler 1994a] O. Reifenschweiler, "Reduced radioactivity of tritium in small titanium particles", _Phys Lett A_ 184:149-153 (3 Jan 1994). OR reports measurements on vacuum-evaporated Ti thin film composed of small Ti particles ~15nm dia. When tritiated to TiT(0.0035) composition, then heated to 275C, radioac- tivity dropped by ~40%, then reincreased to original value when temperature rose to 360C. Another experimental series found that when equal tritium doses were absorbed sequentially on such a Ti thin film, radioactivity did not increase uniformly. OR conjectures an explanation based on nuclear pairing. ------------ [Reifenschweiler 1994b] O. Reifenschweiler, "A more detailed description of our experiments with proposals to improve the experimental technique", unpublished. (March 1994). 33p + 12 figs. Manuscript providing substantially more detail on the experiments described in [Reifenschweiler 1994a]. A few salient points: (1) the higher-concentration experiment mentioned in [R. 1994a] p.150 where the effect "did not occur" is reinterpreted in support of the anomaly. (2) the existence of an oxide layer on the Ti particles is surmised (UHV technique was not used) and considered important to the phenomenon. (3) a preliminary expt with anomalous beta energy distribution is shown. ----------- [Reifenschweiler 1961] O. Reifenschweiler, "Sealed-off neutron tube: the underlying research work", _Philips Res Repts_ 16(5):401-418 (1961). The anomalous activity measurements [R. 1994a, 1994b] were made in connection with the technological development of the Philips neutron-generation tube. That project is described here. No mention of the tritium anomaly. ========================================== Section 2. Titanium Hydride phase diagram ========================================== [Komarek 1983] KL Komarek (ed.), _Titanium: Physico-chemical Properties of its Compounds and Alloys_, Vienna, International Atomic Energy Agency, 1983 (460 pp). Comprehensive review. Values for all the thermodynamic parameters of Ti, TiH, TiD, and TiO compounds. No TiT data. A compilation of measured diffusion rates is also included. -------------------- [McQuillan 1950] AD McQuillan, "An experimental and thermodynamic investigation of the hydrogen-titanium system", _Proc Roy Soc London Ser A_ 204:309-322. Ti:H phase diagram measured with H pressure 0.1-760mmHg, temp 450-950C, exploring alpha-beta and beta-gamma boundaries. Volumetric Sievert technique. Extrapolation to lower temperatures is made. This work does not explore the region of the Reifenschweiler anomaly. -------------- [Gibb 1950] TRP Gibb Jr, & HW Kruschwitz Jr, "The titanium-hydrogen system and titanium hydride. I. Low-pressure studies", _J Amer Chem Soc_ 72:5365-5369 (1950). Ti:H phase diagram measured with H pressure 50-800mmHg, temp 500-800C. The face-centered tetragonal phase (delta) is revealed at full TiH2 stoichiometry. Sievert technique & X-ray crystallography. This work does not explore the region of the Reifenschweiler anomaly. -------------- [Ariyaratnam 1987] SC Ariyatnam, NA Surplice, EH Adem, "Solubility of hydrogen in titanium wires and films at 300K", _J Matls Sci Let_ 6:1349-1350 (1987). Alpha (h.c.p.) to alpha-plus-gamma (f.c.c.) transition is explored in a follow-up to [Kandasamy 1984]. Resistivity measurement is used on a wire. Solubility limit in alpha phase is found to be at 0.0012 H:Ti atomic ratio. The earlier work had found solubility limit of 0.07-0.08 H:Ti in vacuum-evaporated films from the same lot of Ti wire. -------------- [Kandasamy 1984] K Kandasamy, NA Surplice, "The effects of hydrogen sorption on the resistance and work-function of titanium films at 290 K", _J Phys D_ 17:387-398 (1984). High purity Ti was evaporated onto glass under UHV to form films 20-100nm thick and grain sizes 15-400nm. The solubility limit of H in this Ti preparation was 0.07 H:Ti as determined by resistivity peak. Surface area was estimated by Kr physisorption at 78K and low-temp data is interpeted to indicate bulk rather than surface effect for resistivity. Work function is measured with Kelvin probe and correlates with resistivity data. Dissociation pressure of TiH(x) is <1e-8 Torr at 50C for x<1.9. -------------- [Wedler 1966] G Wedler & H Strothenk, "Elektrische und kalorimetrische messungen am system titan/wasserstuff bei 273K", _Zeits Phys Chem NF_ 48:86-101 (1966). In German. Extensive investigation. Ti is evaporated under UHV onto a glass bulb for resistivity and calorimetric measurements. H solubility in alpha Ti is 0.06 to 0.09 in various samples by resistivity peak. The phase diagram *for evaporated films* shows the alpha<->alpha+gamma solvus line projecting almost straight down from the (>320C) alpha-beta line at ~TiH(0.08) (compare [Vitt 1971]), and the alpha+gamma<-> gamma transition (273K) at TiH(0.6) (compare [Korn 1970]). Enthalpies are measured but this reviewer can't read the German-language analysis :-(. -------------- [Paton 1971] NE Paton, BS Hickman, DH Leslie, "Behavior of Hydrogen in alpha- Phase Ti-Al Alloys", Metallurgical Transactions 2:2791-2796 (1971). H solubility in 0.75mm Ti foil is found to follow an Arrhenius law c = a * exp( b/T ) with ~25% hysteresis between heating and cooling. Solubility (cooling curve) is 21ppm by wt (H:Ti = 0.001) at 25C, ~ 10 x higher at 160C. Alloying with Al increases H solubility and hysteresis dramatically. This is interpreted in terms of elastic & plastic strain energy due to expanded lattice of gamma hydride phase. Resistivity measurements and TEM micrographs. -------------- [Vitt 1971] RS Vitt & K Ono, "Hydrogen solubility in alpha titanium", _Metallurg Trans_ 2:608-609 (1971). Solubility in 0.5mm Ti wire is measured by resistivity from 60 to 170C using an age/quench technique. Data is found consistent with [Paton 1971] and some others (but emphatically not compatible with [Lenning 1954], which is deprecated). A composite Arrhenius relationship is fitted for solubility c c = 4.27 * exp (-2.52e3/T) with c in atomic fraction and T in K. Heat of solution of hydride in alpha Ti is 5.0 kcal/mole. ------------------- [Dantzer 1983] P Dantzer, "High temperature thermodynamics of H2 and D2 in titanium, and in dilute titanium oxygen solid solutions", _J Phys Chem Solids_ 44(9):913-923 (1983). A high sensitivity microcalorimeter is used to measure partial molar enthalpy of solution for H in Ti at 440 C. A series of similar expts at ~470 C in alpha TiOy (y=.0047, .012, .018) may be very relevant to Reifenschweiler results. The Ti:O is aged at 1100 C to ensure uniform dissolution of the oxygen. Dantzer plots enthalpy of solution vs H content; it is exothermic and, in Ti, exotherm increases linearly with H conc (up to phase transition at ~9 atom%); this indicates "effective mutual attraction of H atoms in solution". Somewhat similar trend appears in this plot for H in TiOy matrix, and Dantzer fits a straight line to the data. Exotherm is increased ~7% to 10.25 kcal/mol for y=.018, but "mutual attraction" (slope) is reduced compared to pure Ti. Author does not comment but at y=.0047 and .018 the plots show a distinct deviation from linearity with a negative slope at H conc < 2-3 atom%. This reviewer considers this to be evidence of an exothermic interaction between dilute H and O in the lattice which is maximum when x<10at%) are used. H and D solubility is decreased with O concentration. A model in which O at an octohedrally-coordinated site "blocks" H occupancy at the 6 nearest-neighbor tetrahedrally-coordinated sites is found to fit the data well. See also [Yamanaka 1991]. ---------------- [Yamanaka 1991] S Yamanaka, H Ogawa, M Miyake, "Effect of interstitial oxygen on hydrogen solubility in titanium, zirconium, and hafnium", _J Less-Common Metals_ 172-174:85-94 (1991). Data from [Yamanaka 1989] is compared with similar measurements on two other metal oxides. An additional point at TiO(.05) is reported and a relatively sharp change in thermodynamic properties between pure Ti and this composition is suggested by the graphs. ============================================================ Section 3. Surface interactions of Ti, Hydrogen, and Oxygen ============================================================ [Cabrera 1948] N Cabrera & NF Mott, "Theory of the oxidation of metals", _Rept Prog Phys_ 12:163-184 (1948). Seminal paper on the modern theory of metal oxidation. Good treatment of rate laws under varying regimes, Mott potential, etc. Most examples are Cu. ------------------ [Smith 1973] T Smith, "Oxidation of titanium between 25C and 400C", _Surf Sci_ 38:292-312 (1973). The formation and dissolution of oxide surface films is studied experimentally using simultaneous Auger spectroscopy (for superficial layer oxygen content), ellipsometry (oxide layer thickness), and Kelvin probe (surface work function). Activation energy for chemisorption of O- from physisorbed O2 is ~17kcal/mol; heat of adsorption ~7 kcal/mol. Dissolution of O in the Ti bulk (octahedral lattice sites) is significant, esp at elevated temp. For thin films at low oxygen pressure & low temp (Reifenschweiler conditions), cation transport dominates in film growth. Temperature dependence of film growth rate between 24 - 273C, <5nm film thickness, suggests a diffusion activation energy of 4 kcal/mol, unrealistically low. ------------------- [Fromm 1987] E Fromm & H Uchida, "Surface phenomena in hydrogen absorption kinetics of metals and intermetallic compounds", _J Less-Common Met_ 131:1-12 (1987) Gas absorption data for several metals is reviewed with r (probability of an incident molecule's reacting upon impacting the surface) plotted against N (number of gas molecules per unit area already reacted). Ti, like most transition metals exposed to O2, shows an initial plateau at r=1, ascribed to formation of an oxygen monolayer. At high loadings, Ti shows the highest retention of O2 reactivity of all metals plotted. In another presentation, r(H2) is plotted against r(O2) to show "poisoning"; 6 transition metals are very similar and show r(H2) decreased to ~1e-3 while r(O2) remains at 1; then r(H2) and r(O2) fall together. The author contends this shows O2 and H2 absorption are limited at the dissociation step by a single mechanism. Data for hydrogen absorption on Ti through thin metal overlayers (with and without oxide poisoning) is given. ----------------- [Brown 1988] CC Brown & RE Buxbaum, "Kinetics of hydrogen absorption in alpha titanium", _Metallurg Trans A_ 19A:1425ff (1988). A surface energy model which is claimed to model H absorption on bulk Ti from 250 to 500C is given. It contains the following terms: Ec (chemisorption) = 77.4 - 31.4*Theta kJ/mol Hab (absorption) = 48.1 kJ/mol Hbs (bulk-surface activation) = 134.7 kJ/mol No diffusion term is used as experimental data shows a surface-limited phenomenology. ---------------- [Efron 1989] A Efron, Y Lifschitz, I Lewkowicz, MH Mintz, "The kinetics and mechanism of titanium hydride forma- tion", _J Less-Common Met_ 153:23-34 (1989). H was absorbed into vacuum-annealed solid Ti (0.1-1mm thick) over a range of H pressure 30 - 450 torr and temperature 100-250C. Reaction rate was measured (it was found to be independent of H2 pressure) and specimens were examined metallographically. A hydride layer 10-23um thick (low temp) to 90um thick (hi temp) forms and diffusion thru this layer (activation energy ~ 11 kcal/mol) limits reaction rate. Samples with both martensitic & "regular" grain structure were tested with minor differences. ---------------- [Hosoda 1991] N Hosoda, HH Uchida, E Fromm, "Effect of O2, N2, and CO impurities on the H2 absorption rate of Ti films at 300K", _J Less- Common Metals_ 172-174:824-831 (1991). Surface poisoning impedes hydrogen absorption. Extending work reviewed in [Fromm 1987], vacuum-deposited Ti film 30-40nm thick was exposed to several contaminants at varying concentrations and the effects were measured. Oxygen poisoning is examined with a metal oxidation model. ----------------- [Mencer 1991] DE Mencer Jr, TR Hess, T Mebrahtu, DL Cocke, DG Naugle, "Surface reactivity of titanium-aluminum alloys: Ti3Al, TiAl, TiAl3", _J Vac Sci Technol A_ 9(3):1610- 1615 (1991). X-ray photoelectron spectroscopy (XPS) is applied to Ti:Al films codeposited on Cu under UHV, after vacuum and/or O2 treatment at temperatures from 25-600C. Stepwise oxidation of Ti (Ti, TiO2, Ti2O3, TiO2) is revealed thru presentation of XPS spectra for Ti(2p), indicating the corresponding Ti oxidation states (0, 2+, 3+, 4+). Despite the UHV environ- ment (<1e-9 torr), all specimens showed surface oxide; the oxygen was preferentially scavenged as Al2O3 leaving the Ti in metallic state. ---------------- [Azoulay 1992] A Azoulay, U Atzmony, MH Mintz, & N Shamir, "Oxygen-induced surface segregation of hydrogen in polycrystalline titanium", _Surface & Interface Analysis_ 18:643-649 (1992). 0.1 mm Ti foil with slight S contamination was exposed to various levels of oxygen. The rate at which H migrated to the surface was measured with Direct Recoil Spectroscopy (DRS). ============================================================== Section 4. H diffusion in Ti: NMR studies, "hydrogen pairing" ============================================================== [Torrey 1958] HC Torrey, "Effects of translational diffusion on nuclear spin relaxation in simple condensed systems", _Nuovo Cimento, Suppl_ 9(1):95-100 (1958). Diffusion of H in several metal hydrides including gamma TiH(1.77) is measured by pulse(?) NMR at 9.25 MHz. Unlike other metals, Ti shows a double minimum in the T1 relaxation time vs temperature plot, which is unexplained. [Korn 1970] later ascribes it to atmospheric contamination. ------------------ [Korn 1970] C Korn, D Zamir, "NMR study of hydrogen diffusion in the three different phases of the titanium-hydrogen system", _J Phys Chem Solids_ 31:489-502 (1970). 19 MHz pulse NMR was applied to powdered Ti hydrides at temperatures 25-500C. Alpha, beta, gamma, and alpha+{beta|gamma} phases were measured and activation energy for hydrogen diffusion was determined as 11.7 kcal/mol for gamma. The alpha+gamma <-> gamma line is located at TiH(1.50) at 24C and TiH(1.33) at 109C; these give the expected _local_ concentration for gamma precipitates in the alpha+gamma regime. After heavy data-massaging, results from alpha phase showed a minimum relaxation time of ~40 msec near 250C; diffusional activation energy of ~13kcal/mol; H likely located at tetrahedral sites of h.c.p. alpha lattice (but maybe distributed between tetrahedral & ocatahedral sites). Gamma phase signals are much clearer and *do not* replicate [Torrey 1958] except for a single sample which was accidentally oxidized. The oxidized-sample data was never published and has been lost [Korn, priv comm, 1994]. A slight increase in diffusion is adduced for TiH(1.9), T>400C and is conjectured to be caused by thermal excitation of vacancies (compare [Barnes 1987]). The NMR data locates the H at tetrahedral sites in gamma (f.c.c.) hydride, confirming other techniques' results; clustering of H occupancy in the lattice is noted for H:Ti < 1.6. ------------------- [Wasilewski 1954] RJ Wasilewski & GM Kehl, _Mettalurgia_ 50:225 (1954). Cited by [Korn 1970] for a determination of H-in-alpha-Ti diffusion activation energy = 12.4 kcal/mol. This reviewer has not read the paper. Compare [Efron 1989] where diffusion thru gamma layer (E=11 kcal/mol) is considered the rate-limiting step. ------------------- [Seymour 1982] EFW Seymour, "NMR studies of hydrogen diffusion in metal hydrides", _J Less-Common Metals_ 88:323-334 (1982). ES reviews metal-hydride NMR and finds several types of reported theory:experiment mismatch much reduced with recently refined theretical methods. Data of [Torrey 1958] is presented and it is proposed that the lower-temperature minimum in relaxation time is caused by paramagnetic impurities. Data from YH(1.98) with small admixture of Gd is shown to illustrate the effect of such impurities. This effect is marked but not as striking as the [Torrey 1958] data. A reverse isotope shift, with the T1 minimum occuring at a lower temperature in TiD(1.55) than in the hydride, is related to the electric quadrupole moment of D (and therefore should not be expected for T in place of D). ------------------- [Barnes 1987] RG Barnes et al, "Dynamical evidence of hydrogen sublattice melting in metal-hydrogen systems", _J Less-Common Metals_ 129:279-285 (1987). The first of a series of papers by a group led by Barnes & Torgenson, pursuing a high-temperature anomaly in the T1 NMR relaxation time for several metal hydrides. A parallel is drawn to fast-ion (superionic) conductors such as PbF2; correlated motion of H packets is proposed along with the formation of interstitial H clusters. Experiments were performed with Sc, Y, La, Ti, and Zr hydrides and deuterides, but no Ti data is reported here or in the subsequent papers this reviewer has located. ------------------- [Barnes 1989] R Barnes, "NMR studies of hydrogen in metals and hydrides at very high temperatures", _Zeits Phys Chem NF_ 164:841-851 (1989). An amplification of [Barnes 1987]. Data from Nb(0.75)V(0.25)H(0.23) is shown to illusrate the "typical" behavior that hydrogen diffusion itself increases smoothly with temperature despite the anomalous "turn-down" in T1 relaxation time at high temperatures. Measurements in a Sc lattice, whose 45Sc nuclei are accessible to NMR, shows the same anomaly. A single explanation is sought which applies for all lattices (f.c.c., b.c.c., h.c.p.) in which the anomaly has been seen, at high and low H concentration. A two-state model, in which H atoms alternate between a mobile state and an immobile state, staying longer in the immobile state, is evaluated and found to be plausible. ------------------- [Han 1989] J-W Han, LR Lichty, DR Torgeson, EFW Seymour, RG Barnes, "Anomalous proton spin-lattice relaxation at high temperatures in bcc transition-metal - hydrogen solid-solution systems", _Phys Rev B_ 40(13):9025-9034 (1989). A refinement of [Barnes 1989], the two-state model is elucidated and it is found that the immobile state "resting time" must increase exponentially with temperature: the immobile state behaves as an excited state with an activation energy ~0.71 eV for an NbVH system. This is considered counterintuitive as no mechanism seems at hand. -------------------- [Han 1991] J-W Han, DR Torgeson, RG Barnes, DT Peterson, "Anomalous proton-spin - latice relaxation at high temperatures in zirconium dihydrides", _Phys Rev B_ 44(22):12353ff (1991). The two-state model is applied to data from ZrH(1.58-1.98). -------------------- [Baker 1992] DB Baker, N Adolphi, MS Conradie, PA Fedders, RE Norberg, RG Barnes, DR Torgeson, "Evidence for the high-temperature spin-relaxation anomaly in metal hydrides", _Phys Rev B_ 46(1):184-187 (1992). NMR measurements are extended to 341 MHz on a Nb(0.5)V(0.5)H(0.36) sample and a new frequency dependence is observed in the anomaly. The data imply an excited immobile state with a large second moment. Interstitial hydrogen molecules are proposed. -------------------- [Cotts 1991] RM Cotts, "Would hydrogen pairing at high temperatures account for NMR relaxation anomalies in metal hydrides?", _J Less-Common Metals_ 172-174:467-474 (1991). RC reviews the data from the Barnes-Torgeson group and finds the data difficult to explain. The idea of "close pairs" or H2 molecules occuring in the hydride lattice is considered, and surprisingly close proximity is required in order for the H-H interaction to be strong enough to explain the data. For ScH(1.83), an H-H spacing of no more than 0.75 Angstr (the same as H2 gas molecule) is necessary. In Nb(0.75)V(0.25)H(0.23), the spacing must be less than 0.39 Angstr! This seems to be a far fetch, but no less attractive than other proposed explanations. -------------------- [Baker 1994] DB Baker, MS Conradi, RE Norberg, RG Barnes, DR Torgeson, "Explanation of the high-temperature relaxation anomaly in a metal-hydrogen system", _Phys Rev B_ 49(17):11773- 11782 (May 1994). The interstitial molecule hypothesis [Baker 1992] is abandoned based on new evidence (especially the disappearance of the effect under inert gas overpressure) in favor of a surface phenomenon. A model based on rapid exchange of H between the hydride and molecular gas is proposed and fitted to data in the Nb(0.5)V(0.5)H(0.36) system. An appendix considers a model based on fast-relaxing surface H states (atoms or H(2)- ions) but finds these specific mechanisms unable to explain the isotope dependency (anomalous relaxation rate is reduced by ~4.3X when 90% of the H is replaced by D). The authors are concerned that the qualitatively similar 45Sc anomaly [Barnes 1989] remains unexplained by the gas-exchange model. Further anecdotal evidence of surface phenomenology is "poisoning" by oxygen (when packed with MgO) and nitrogen (from BN). ================================= Section 5. Tritium-specific data ================================= [Bach 1980] P Bach, "Improvement of the room-temperature behavior of metal tritides with respect to 3He release: Titanium", _Appl Phys Lett 37(5):492-494 (1 Sept 1980). Thin films (0.6 mg/cm^2) [ ~ 15 nm thick] were tritided to TiT(1.19-1.72) composition after HV or UHV outgassing. 3He release was monitored over extended (up to 6 yrs) aging. The author does not comment, but this reviewer notes that, in contrast to a 1975 Oak Ridge study by Perkins et al, _none_ of the thin-film samples achieved ultimate helium evolution rates as high as predicted. For three different preparation techniques, the maximum reported values of v (desorption rate / generation rate) were 20%, 25%, and 40%. This is consistent with anomalously low tritium activity under these conditions. --------------------- [Schober 1989] T Schober & K Farrell, "Helium bubbles in alpha-Ti and Ti tritide arising from tritium decay: a TEM study", _J Nucl Mat_ 168:171-177 (1989). Solid Ti disks were soaked at 600C in tritium gas for several weeks. T was dispersed in the alpha phase and its decay gave rise to 3He bubbles. Subsequently the material was cooled to room temperature and sectioned for microscopic analysis after verious periods of aging. Many interesting features about the specimens are reported, but few conclusions can be drawn. During cooling, gamma(f.c.c.) tritide (here called delta) precipitated in platelets 20-100nm thick, embedded in alpha (h.c.p.) matrix. In the alpha matrix, one might expect uniform bubble concentration, but in fact it was "patchy". As expected, residual T concentration in matrix was very low and no new He appeared to form there. In the tritide platelets, tight clusters of high-density bubbles formed, and they grew with time as expected (volume proportional to time) without any apparent increase in the number of bubbles. However, an approximate analysis showed the total volume of 3He was substantially smaller than expected from the assumed platelet composition TiT(1.5). Many platelets were internally divided by a boundary separating a bubble-free zone from a populated zone. Strain fields at the bubbles were investigated. -------------------- [Jordan 1967] KC Jordan, BC Blanke, WA Dudley, "Half-life of tritium", _J Inorg Nucl Chem_ 29:2129-2131 (1967). A brief report of the calorimetric determination of tritium half-life to high precision (12.346 +/-0.002 yr). A Ti tritide sample of unknown composition was monitored for heat output over a period of 6.5 years; the output showed exponential decay as expected. However deviations from this curve were found during the first few months after fabrication. After ~4 yrs an experimental heat treatment at 90C caused an anomalous drop in output of ~0.05% (5 sigma - quite detectable!) for 6 months, followed by a spontaneous burst of high output, then a return to the expected curve for ~ 2 more years. No explanation is proposed. -------------------- [Forker 1989] M Forker, W Herz, D Simon, R Lasser, "TDPAC study of tritium diffusion in the hafnium-tritium system", _Zeits Phys Chem NF_ 164:889-894 (1989). Time Differential Perturbed Angular Correlation (TDPAC) measurements were made to monitor diffusion characteristics of tritium in HfT(1.9). A normal relaxation-time minimum was found near 275C, corresponding to a diffusional activation energy of ~0.36 eV. Between 20C and 125C, an anomalous increase of vacancy concentration appeared. The authors propose desorption of T from the sample, but find this unconvincing and suggest further studies. The anomalous effect is reversible with temperature cycling and has not occurred with Hf:H samples. -------------------- [Budick 1992] B Budick, J Chen, H Lin, "Bremsstrahlung from tritium beta decay", _Phys Rev C_ 46(4):1267-1275 (Oct 1992). Investigation of the soft x-ray bremsstrahlung spectrum from T decay, including the effect of surrounding matter on external bremsstrahlung (EB) versus internal bremsstrahlung (IB) contri- butions. Extensive experimentation with T(2) - noble-gas mixtures. GM-tube measurements in [Reifenschweiler 1994a] were based on bremsstrahlung x-rays.