Anomalous temperature dependence of the order parameter of a superconductor with weakly correlated impurities

It is shown that weak correlations between pair-breaking impurities in superconductors influence the temperature dependence of the order parameter within the Ginzburg-Landau region if correlation radius of impurities R is greater than the coherence length of the superconductor ξ₀. The dependence of the square of the average order parameter on a temperature difference T _c − T changes its slope in a region $ \xi _0 \sqrt {{{T_c } \mathord{\left/ {\vphantom {{T_c } {\left( {T_c - T} \right)}}} \right. \kern-\nulldelimiterspace} {\left( {T_c - T} \right)}}} \sim R $ \xi _0 \sqrt {{{T_c } \mathord{\left/ {\vphantom {{T_c } {\left( {T_c - T} \right)}}} \right. \kern-\nulldelimiterspace} {\left( {T_c - T} \right)}}} \sim R . The influence of correlations of impurities on other thermodynamic properties of superconductors is discussed.     

Structures of distorted phases and critical and noncritical atomic displacements of elpasolite Rb<Subscript>2</Subscript>KInF<Subscript>6</Subscript> during phase transitions

The structures of all three phases of the Rb₂KInF₆ crystal have been determined from the experimental X-ray diffraction data for the powder sample. The refinement of the profile and structural parameters has been carried out by the technique implemented in the DDM program, which minimizes the differences between the derivatives of the calculated and measured X-ray intensities over the entire profile of the X-ray diffraction pattern. The results obtained have been discussed using the group-theoretical analysis of the complete order-parameter condensate, which takes into account the critical and noncritical atomic displacements and permits the interpretation of the experimental data obtained previously. It has been reliably established that the sequence of changes in the symmetry during phase transitions in Rb₂KInF₆ can be represented as $ Fm\bar 3m\xrightarrow[{0,0,\phi }]{{11 - 9\left( {\Gamma _4^ + } \right)}}{{I114} \mathord{\left/ {\vphantom {{I114} {m\xrightarrow[{\left( {\psi ,\phi ,\phi } \right)}]{{11 - 9\left( {\Gamma _4^ + } \right) \oplus 10 - 3\left( {X_3^ + } \right)}}{{P12_1 } \mathord{\left/ {\vphantom {{P12_1 } {n1}}} \right. \kern-\nulldelimiterspace} {n1}}}}} \right. \kern-\nulldelimiterspace} {m\xrightarrow[{\left( {\psi ,\phi ,\phi } \right)}]{{11 - 9\left( {\Gamma _4^ + } \right) \oplus 10 - 3\left( {X_3^ + } \right)}}{{P12_1 } \mathord{\left/ {\vphantom {{P12_1 } {n1}}} \right. \kern-\nulldelimiterspace} {n1}}}} $ Fm\bar 3m\xrightarrow[{0,0,\phi }]{{11 - 9\left( {\Gamma _4^ + } \right)}}{{I114} \mathord{\left/ {\vphantom {{I114} {m\xrightarrow[{\left( {\psi ,\phi ,\phi } \right)}]{{11 - 9\left( {\Gamma _4^ + } \right) \oplus 10 - 3\left( {X_3^ + } \right)}}{{P12_1 } \mathord{\left/ {\vphantom {{P12_1 } {n1}}} \right. \kern-\nulldelimiterspace} {n1}}}}} \right. \kern-\nulldelimiterspace} {m\xrightarrow[{\left( {\psi ,\phi ,\phi } \right)}]{{11 - 9\left( {\Gamma _4^ + } \right) \oplus 10 - 3\left( {X_3^ + } \right)}}{{P12_1 } \mathord{\left/ {\vphantom {{P12_1 } {n1}}} \right. \kern-\nulldelimiterspace} {n1}}}} .     

New criterion for comparison of classical theory of nucleation in superheated liquids with experimental data

We have obtained inequality $ 1 - {{\Delta \bar \tau } \mathord{\left/ {\vphantom {{\Delta \bar \tau } {\bar \tau }}} \right. \kern-\nulldelimiterspace} {\bar \tau }} < \left( {J \cdot V \cdot \bar \tau } \right)^{ - 1} < 1 + {{\Delta \bar \tau } \mathord{\left/ {\vphantom {{\Delta \bar \tau } {\bar \tau }}} \right. \kern-\nulldelimiterspace} {\bar \tau }} $ 1 - {{\Delta \bar \tau } \mathord{\left/ {\vphantom {{\Delta \bar \tau } {\bar \tau }}} \right. \kern-\nulldelimiterspace} {\bar \tau }} < \left( {J \cdot V \cdot \bar \tau } \right)^{ - 1} < 1 + {{\Delta \bar \tau } \mathord{\left/ {\vphantom {{\Delta \bar \tau } {\bar \tau }}} \right. \kern-\nulldelimiterspace} {\bar \tau }} , where J is the frequency of homogeneous nucleation, V and $ \bar \tau $ \bar \tau are, respectively, volume and average lifetime of the superheated liquid, and $ {{\Delta \bar \tau } \mathord{\left/ {\vphantom {{\Delta \bar \tau } {\bar \tau }}} \right. \kern-\nulldelimiterspace} {\bar \tau }} $ {{\Delta \bar \tau } \mathord{\left/ {\vphantom {{\Delta \bar \tau } {\bar \tau }}} \right. \kern-\nulldelimiterspace} {\bar \tau }} is relative statistical error $ \bar \tau $ \bar \tau . Inequality appears to be a consequence of nucleation homogeneity and stability used at its deduction and taken in the theory as initial and determinant assumption. Calculations with the use of experimental data for the boundaries of the attainable superheating show that inequality is not satisfied. Thus, experimental data can not be considered a proof of the theory fundamentals.     

Self-diffusion in α-Fe2O3 natural single crystals

Measurements of¹⁸O self-diffusion in hematite (Fe₂O₃) natural single crystals have been carried out as a function of temperature at constant partial pressure a_{O 2}=6.5·10^?2 in the temperature range 890 to 1227 °C. The a_{O 2} dependence of the oxygen self-diffusion coefficient at fixed temperature T=1150 °C has also been deduced in the a_{O 2} range 4.5·10^?4 - 6.5·10^?1. The concentration profiles were established by secondary-ion mass spectrometry; several profiles exhibit curvatures or long tails; volume diffusion coefficients were computed from the first part of the profiles using a solution taking into account the evaporation and the exchange at the surface. The results are well described by $$D_O \left( {{{cm^2 } \mathord{\left/ {\vphantom {{cm^2 } s}} \right. \kern-\nulldelimiterspace} s}} \right) = 2.7 \cdot 10^8 a_{O_2 }^{ - 0.26} \exp \left( { - \frac{{542\left( {{{kJ} \mathord{\left/ {\vphantom {{kJ} {mol}}} \right. \kern-\nulldelimiterspace} {mol}}} \right)}}{{RT}}} \right)$$ From fitting a grain boundary diffusion solution to the profile tails, the oxygen self-diffusion coefficient in sub-boundaries has been deduced. They are well described by $$D''_O \left( {{{cm^2 } \mathord{\left/ {\vphantom {{cm^2 } s}} \right. \kern-\nulldelimiterspace} s}} \right) = 3.2 \cdot 10^{25} a_{O_2 }^{ - 0.4} \exp \left( { - \frac{{911\left( {{{kJ} \mathord{\left/ {\vphantom {{kJ} {mol}}} \right. \kern-\nulldelimiterspace} {mol}}} \right)}}{{RT}}} \right)$$ Experiments performed introducing simultaneously¹⁸O and⁵⁷Fe provided comparative values of the self-diffusion coefficients in volume: iron is slower than oxygen in this system showing that the concentrations of atomic point defects in the iron sublattice are lower than the concentrations of atomic point defects in the oxygen sublattice. The iron self-diffusion values obtained at T>940 °C can be described by $$D_{Fe} \left( {{{cm^2 } \mathord{\left/ {\vphantom {{cm^2 } s}} \right. \kern-\nulldelimiterspace} s}} \right) = 9.2 \cdot 10^{10} a_{O_2 }^{ - 0.56} \exp \left( { - \frac{{578\left( {{{kJ} \mathord{\left/ {\vphantom {{kJ} {mol}}} \right. \kern-\nulldelimiterspace} {mol}}} \right)}}{{RT}}} \right)$$ The exponent - 1/4 observed for the oxygen activity dependence of the oxygen self-diffusion in the bulk has been interpreted considering that singly charged oxygen vacancies V _O ^? are involved in the oxygen diffusion mechanism. Oxygen activity dependence of iron self-diffusion is not known accurately but the best agreement with the point defect population model is obtained considering that iron self-diffusion occurs both via neutral interstitals Fe _x ⁱ and charged ones.     

Superconductive superheating field for finiteκ

We have calculated analytically the superheating fieldH _sh for bulk superconductors, correct to second order in. We find , which agrees well with numerical computations for<0.5. The surface order parameter is , and the penetration depth is .     

Der Einfluß der Wärmebewegung auf die Breite einer Mössbauerlinie über den Dopplereffekt zweiter Ordnung

AtT=0 a perfect Mössbauer line has natural line widthΓ=?/τ _n . However, with rising temperature the width increases. The reason of the line broadening is the second order Doppler effect which causes a stochastic frequency modulation of theγ-radiation, reflecting the thermal motion of the Mössbauer atom. Following Josephson in treating the second order Doppler shift as a mass changeΔM=E _n/c² of theγ-emitting atom caused by the loss of nuclear excitation energy E _n , and using the well known relaxation formalism for calculating theγ-frequency spectrum, the line broadeningΔ Γ is evaluated within the framework of harmonic lattice theory. For a parabolic lattice frequency spectrum with Debye-temperature Θ one obtains $$\Delta {\Gamma \mathord{\left/ {\vphantom {\Gamma \Gamma }} \right. \kern-\nulldelimiterspace} \Gamma } = \left( {{{\tau _n } \mathord{\left/ {\vphantom {{\tau _n } {\tau _c }}} \right. \kern-\nulldelimiterspace} {\tau _c }}} \right) \cdot \left( {{{E_n } \mathord{\left/ {\vphantom {{E_n } {Mc^2 }}} \right. \kern-\nulldelimiterspace} {Mc^2 }}} \right) \cdot F\left( {{T \mathord{\left/ {\vphantom {T \Theta }} \right. \kern-\nulldelimiterspace} \Theta }} \right),where\tau _c = {{\rlap{--} h} \mathord{\left/ {\vphantom {{\rlap{--} h} k}} \right. \kern-\nulldelimiterspace} k}\Theta $$ is the correlation time of the lattice vibrations. The functionF(T/Θ) may be expanded in powers ofT/Θ, yielding $$F\left( {{T \mathord{\left/ {\vphantom {T \Theta }} \right. \kern-\nulldelimiterspace} \Theta }} \right) = 9720\pi \left( {{T \mathord{\left/ {\vphantom {T \Theta }} \right. \kern-\nulldelimiterspace} \Theta }} \right)^7 forT<< \Theta $$ and $$F\left( {{T \mathord{\left/ {\vphantom {T \Theta }} \right. \kern-\nulldelimiterspace} \Theta }} \right) = 2.7\pi \left( {{T \mathord{\left/ {\vphantom {T \Theta }} \right. \kern-\nulldelimiterspace} \Theta }} \right)^2 forT > > \Theta $$ , respectively. Although unavoidable, the line broadening is obviously too small to be observable by means of the present experimental technique.     

Weber potential from finite velocity of action?

The Weber potential energy U for charges q and q' separated by the distance R is U = (qq'/R)[1 – (dR/dt)²/2c²]. If this potential arises from a finite velocity c of energy transfer Q', where the retarded rate of transfer from q' to q is dQ(t-R/c)/dt = Q'[1 – (dR/dt)/c] and where the advanced rate from q to q' is dQ(t+R/c)/dt = Q'[1 + (dR/dt)/c], then the resultant time-average root-mean-square action is given by . Identifying Q' with the Coulomb potential energy qq'/R, the Weber potential is obtained. Using the same argument, Newtonian gravitation yields a corresponding Weber potential energy, qq'/R being replaced by ( - Gmm'/R).     

Comment on “Evaluation of concentration-depth profiles by sputtering in SIMS and AES” by S. Hofmann

The statistics of the sputtering process, which has been used to explain sputterbroadening effect due to surface roughness, has been treated with conditional probabilities. This results in the relationship, , instead of derived by S. Hofmann [Appl. Phys.9, 59 (1976)], where δz,z, and are the depth resolution, sputtered depth and sputtering yield, respectively.     

Neutral meson oscillations and equivalence principle for particles and antiparticles

Oscillations of neutral meson (K ⁰-$ \overline {K^0 } $ \overline {K^0 } , D ⁰-$ \overline {D^0 } $ \overline {D^0 } , and B ⁰-$ \overline {B^0 } $ \overline {B^0 } are extremely sensitive to the meson and antimeson energies at rest. This energy is determined as mc ²—with the corresponding inertial mass—and as the energy of gravitational interaction. Assuming that the CPT theorem is correct for inertial masses and estimating the gravitational potential for which the largest contribution originates from the field of the galaxy center, we obtain the estimate from experimental data on K ⁰-$ \overline {K^0 } $ \overline {K^0 } oscillations:

Pre-exponential factor of the hopping conductivity in disordered carbon films

The conductivity of carbon films grown by polymethylphenylsiloxane vapor decomposition in stimulated dc discharge plasma was studied. It is found that the Mott hopping conductivity $ \sigma \left( T \right) = \sigma _0 \left( T \right)\exp \left\{ { - \frac{{T_0 }} {T}^{{1 \mathord{\left/ {\vphantom {1 4}} \right. \kern-\nulldelimiterspace} 4}} } \right\} $ \sigma \left( T \right) = \sigma _0 \left( T \right)\exp \left\{ { - \frac{{T_0 }} {T}^{{1 \mathord{\left/ {\vphantom {1 4}} \right. \kern-\nulldelimiterspace} 4}} } \right\} is characteristic of the samples under study in the temperature range of 80–400 K in the electric field E to 5 · 10⁴ V/cm. An analysis of the pre-exponential factor σ ₀(T) = σ ₀₀(T ₀)T ^α allowed the conclusion that the hopping transport is most adequately described in the model with the exponential energy dependence of the density of localized states for which α = −1/2 and the universal relation ln σ ₀₀ −T ₀^1/4 0 is valid, which is satisfied in the range where the parameter σ ₀₀ varies by eight orders of magnitude.     

Diffusion of proton in alumina-rich nonstoichiometric magnesium aluminate spinel

The phenomenon of the diffusion of proton and deuteron in a single crystal of magnesium aluminate spinel was studied by infrared absorption. The chemical diffusion coefficient of proton was determined by the relaxation time of the absorption intensity upon the substitution of deuteron with proton. The temperature dependence of the chemical diffusion coefficient of proton for was expressed as . The chemical diffusion coefficient of proton was found to be independent of the composition of spinel and of the atmosphere. Paper presented at the 11th Euro Conference on the Science and Technology of Ionics, Batz-sur-Mer, Sept. 9–15 2007.     

“Magic numbers” in the melting of a cluster of point charges on a sphere

The thermodynamic properties of a cluster of point Coulomb charges on a sphere have been analyzed using the Monte Carlo method for the number of charges 20 ≤ N ≤ 90. The ground state of the system of charges is described in the model of a closed quasi-two-dimensional triangular lattice with topological defects. We have determined the dependence of the Lindeman parameter δ_L of this system on N and on the dimensionless parameter $ \tilde T $ \tilde T , which is proportional to the temperature T and to the radius R of the cluster: $ \tilde T = {{k_B T\varepsilon R} \mathord{\left/ {\vphantom {{k_B T\varepsilon R} {e^2 }}} \right. \kern-\nulldelimiterspace} {e^2 }} $ \tilde T = {{k_B T\varepsilon R} \mathord{\left/ {\vphantom {{k_B T\varepsilon R} {e^2 }}} \right. \kern-\nulldelimiterspace} {e^2 }} , where ∈ is the dielectric constant of the medium and e is the charge of a particle. The “magic numbers,” i.e., the N values, for which the melting point of the closed triangular lattice of charges is much higher than those for neighboring N values, have been found. The evolution of the lattice-melting mechanisms with an increase in the number of charges N in a mesoscopic cluster has been analyzed. For N ≤ 32, the melting of the lattice does not involve dislocations (nontopological melting); this behavior of the mesoscopic system of charges on the sphere differs from the behavior of the extended planar two-dimensional system. At N ≳ 50, melting is accompanied by the formation of dislocations. The mechanism of dislocation-free non-topological melting of a closed lattice, which occurs at small N values and is associated with the cooperative rotational motion of “rings” of particles, has been analyzed. The model has various implementations in the mesoscopic region; in particular, it describes the system of electrons over the liquid-helium cluster, the liquid-helium cluster with incorporated charged particles, a multielectron bubble in liquid helium, a charged quantum dot, etc.     

Asymptotic behavior of Mayer cluster sums for the one-dimensional Ising model

The properties of the high-field polynomialsL _n (u) for the one-dimensional spin 1/2 Ising model are investigated. [The polynomialsL _n (u) are essentially lattice gas analogues of the Mayer cluster integralsb _n (T) for a continuum gas.] It is shown thatu ^?1 L _n (u) can be expressed in terms of a shifted Jacobi polynomial of degreen?1. From this result it follows thatu ^?1 L _n (u); n=1, 2,... is a set of orthogonal polynomials in the interval (0, 1) with a weight functionω(u)=u, andu ^?1 L _n (u) hasn?1 simple zerosu _n (v); v=1, 2,...,n?1 which all lie in the interval 0<u<1. Next the detailed behavior ofL _n (u) asn→∞ is studied. In particular, various asymptotic expansions forL _n (u) are derived which areuniformly valid in the intervalsu<0, 0<u<1, andu>1. These expansions are then used to analyze the asymptotic properties of the zeros {u _n (v); v=1, 2,...,n?1}. It is found that $$\begin{array}{*{20}c} {u_n (v) \sim \tfrac{1}{4}({{j_{1,v} } \mathord{\left/ {\vphantom {{j_{1,v} } n}} \right. \kern-\nulldelimiterspace} n})^2 [1 - ({{j_{1,v}^2 } \mathord{\left/ {\vphantom {{j_{1,v}^2 } {12}}} \right. \kern-\nulldelimiterspace} {12}})n^{ - 1} + ({{j_{1,v}^2 } \mathord{\left/ {\vphantom {{j_{1,v}^2 } {700)( - 3 + 2j_{1,v}^2 )n^{ - 4} }}} \right. \kern-\nulldelimiterspace} {700)( - 3 + 2j_{1,v}^2 )n^{ - 4} }}} \\ { + ({{j_{1,v}^2 } \mathord{\left/ {\vphantom {{j_{1,v}^2 } {20160)(40 + 4j_{1,v}^2 - j_{1,v}^4 }}} \right. \kern-\nulldelimiterspace} {20160)(40 + 4j_{1,v}^2 - j_{1,v}^4 }})n^{ - 6} + \cdot \cdot \cdot ]} \\ {u_n (n - v) \sim 1 - ({{j_{0,v}^2 } \mathord{\left/ {\vphantom {{j_{0,v}^2 } 4}} \right. \kern-\nulldelimiterspace} 4})n^{ - 2} + ({{j_{0,v}^2 } \mathord{\left/ {\vphantom {{j_{0,v}^2 } {48)( - 2 + j_{0,v}^2 )n^{ - 4} }}} \right. \kern-\nulldelimiterspace} {48)( - 2 + j_{0,v}^2 )n^{ - 4} }}} \\ { + ({{j_{0,v}^2 } \mathord{\left/ {\vphantom {{j_{0,v}^2 } {2880)(2 + 9j_{0,v}^2 - 2j_{0,v}^4 )n^{ - 6} + \cdot \cdot \cdot }}} \right. \kern-\nulldelimiterspace} {2880)(2 + 9j_{0,v}^2 - 2j_{0,v}^4 )n^{ - 6} + \cdot \cdot \cdot }}} \\ \end{array} $$ asn→∞v fixed, wherej _k,v denotes thevth zero of the Bessel functionJ _k(^z)     

Search for double electron capture of 106Cd

A search for double electron capture of ¹⁰⁶Cd was performed at the Modane Underground Laboratory (4800 m w.e.) using a low-background and high-sensitivity multidetector spectrometer TGV-2 (Telescope Germanium Vertical). New limits on β ⁺/EC, EC/EC decays of ¹⁰⁶Cd were obtained from preliminary calculations of experimental data accumulated for 4800 h of measurement of 10 g of ¹⁰⁶Cd with enrichment of 75%. They are > 9.1 × 10¹⁸ yr, > 1.9 × 10¹⁹ yr for transitions to the first 2⁺, 511.9 keV excited state of ¹⁰⁶Pd, and > 1.3 × 10¹⁹ yr, > 6.2 × 10¹⁹ yr for transitions to the ground 0⁺ state of ¹⁰⁶Pd. All limits are given at 90% C.L. The text was submitted by the authors in English.     

Influence of a strong longitudinal static electric field on free carrier faraday effect in an n-type InSb at room temperature at submillimeter wave frequencies

The expression for free carrier Faraday rotation and for ellipticity , as the function of the applied parallel static electric field and static magnetic field for a given value of wave angular frequency and electron concentration N₀, are obtained and theoretically analyzed with the aid of one-dimensional linearized wave theory and Kane's non-parabolic isotropic dispersion law. It is shown that the maximum Faraday rotation occurs near the cyclotron resonance condition, which can be expressed as , where , , and . Here m^* and e denote the effective mass and charge of electron, respectively. _g is the forbidden bandgap of semiconductor. v₀ is the carrier drift velocity, which is a non-linear function of E₀ in high field condition. A possibility of a simple way of determining the non-linear v₀ vs E₀ characteristics of semiconductors by the measurement of Faraday rotation is also discussed.     

Study of the mechanisms of pre-equilibrium nuclear reactions in the microscopic approach

The mechanisms of pre-equilibrium nuclear reactions are investigated within the Statistical Multistep Direct Process (SMDP) + Statistical Multistep Compound Process (SMCP) formalism. It has been shown that from an analysis of linear part in such dependences as $$\ln \left[ {{{\frac{{d^2 \sigma }}{{d\varepsilon _b d\Omega _b }}} \mathord{\left/ {\vphantom {{\frac{{d^2 \sigma }}{{d\varepsilon _b d\Omega _b }}} {\varepsilon _b^{1/2} }}} \right. \kern-\nulldelimiterspace} {\varepsilon _b^{1/2} }}} \right]upon\varepsilon _b $$ and $$\ln \left[ {{{\frac{{d\sigma ^{SMDP \to SMCP} }}{{d\varepsilon _b }}} \mathord{\left/ {\vphantom {{\frac{{d\sigma ^{SMDP \to SMCP} }}{{d\varepsilon _b }}} {\frac{{d\sigma ^{SMDP} }}{{d\varepsilon _b }}}}} \right. \kern-\nulldelimiterspace} {\frac{{d\sigma ^{SMDP} }}{{d\varepsilon _b }}}}} \right]upon{{U_B } \mathord{\left/ {\vphantom {{U_B } {\left( {E_a - B_b } \right)}}} \right. \kern-\nulldelimiterspace} {\left( {E_a - B_b } \right)}}$$ one can extract information about the type of mechanism (SMDP, SMCP, SMDP→SMCP) and the number of stages of the multistep emission of secondary particles. In the above approach, we have discussed the experimental data for a broad class of reactions in various entrance and exit channels.     

O(G μ 2 m t 4 ) electroweak radiative corrections to theZb \bar b vertex

The leading heavy-top two-loop corrections to theZb \(\bar b\) vertex are determined from a direct evaluation of the corresponding Feynman diagrams in the largem _t limit. The leading one-loop top-mass effect is enhanced by \([{{1 + G_\mu m_t^2 ({{9 - \pi ^2 } \mathord{\left/ {\vphantom {{9 - \pi ^2 } 3}} \right. \kern-0em} 3})} \mathord{\left/ {\vphantom {{1 + G_\mu m_t^2 ({{9 - \pi ^2 } \mathord{\left/ {\vphantom {{9 - \pi ^2 } 3}} \right. \kern-0em} 3})} {(8\pi ^2 \sqrt 2 )}}} \right. \kern-0em} {(8\pi ^2 \sqrt 2 )}}]\) . Our calculation confirms a recent result of Barbieri et al..     

Light quark masses in quantum chromodynamics and chiral symmetry breaking

We derive model independent lower bounds for the sums of effective quark masses \(\bar m_u + \bar m_d \) and \(\bar m_u + \bar m_s \) . The bounds follow from the combination of the spectral representation properties of the hadronic axial currents two-point functions and their behavior in the deep euclidean region (known from a perturbative QCD calculation to two loops and the leading non-perturbative contribution). The bounds incorporate PCAC in the Nambu-Goldstone version. If we define the invariant masses \(\hat m\) by $$\bar m_i = \hat m_i \left( {{{\frac{1}{2}\log Q^2 } \mathord{\left/ {\vphantom {{\frac{1}{2}\log Q^2 } {\Lambda ^2 }}} \right. \kern-\nulldelimiterspace} {\Lambda ^2 }}} \right)^{{{\gamma _1 } \mathord{\left/ {\vphantom {{\gamma _1 } {\beta _1 }}} \right. \kern-\nulldelimiterspace} {\beta _1 }}} $$ and <F ²> is the vacuum expectation value of $$F^2 = \Sigma _a F_{(a)}^{\mu v} F_{\mu v(a)} $$ , we find, e.g., $$\hat m_u + \hat m_d \geqq \sqrt {\frac{{2\pi }}{3} \cdot \frac{{8f_\pi m_\pi ^2 }}{{3\left\langle {\alpha _s F^2 } \right\rangle ^{{1 \mathord{\left/ {\vphantom {1 2}} \right. \kern-\nulldelimiterspace} 2}} }}} $$ ; with the value <α _u F ²?0.04GeV⁴, recently suggested by various analysis, this gives $$\hat m_u + \hat m_d \geqq 35MeV$$ . The corresponding bounds on \(\bar m_u + \bar m_s \) are obtained replacingm _π ² f _π bym _K ² f _K . The PCAC relation can be inverted, and we get upper bounds on the spontaneous masses, \(\hat \mu \) : $$\hat \mu \leqq 170MeV$$ where \(\hat \mu \) is defined by $$\left\langle {\bar \psi \psi } \right\rangle \left( {Q^2 } \right) = \left( {{{\frac{1}{2}\log Q^2 } \mathord{\left/ {\vphantom {{\frac{1}{2}\log Q^2 } {\Lambda ^2 }}} \right. \kern-\nulldelimiterspace} {\Lambda ^2 }}} \right)^d \hat \mu ^3 ,d = {{12} \mathord{\left/ {\vphantom {{12} {\left( {33 - 2n_f } \right)}}} \right. \kern-\nulldelimiterspace} {\left( {33 - 2n_f } \right)}}$$ .