* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download 1 Derivation of Schrödinger`s equation Mikhail Batanov, Associate
Copenhagen interpretation wikipedia , lookup
Lattice Boltzmann methods wikipedia , lookup
Noether's theorem wikipedia , lookup
Scalar field theory wikipedia , lookup
Renormalization wikipedia , lookup
Quantum electrodynamics wikipedia , lookup
Symmetry in quantum mechanics wikipedia , lookup
Bohr–Einstein debates wikipedia , lookup
History of quantum field theory wikipedia , lookup
Probability amplitude wikipedia , lookup
Molecular Hamiltonian wikipedia , lookup
Atomic theory wikipedia , lookup
Double-slit experiment wikipedia , lookup
Identical particles wikipedia , lookup
Electron scattering wikipedia , lookup
Hidden variable theory wikipedia , lookup
Erwin Schrödinger wikipedia , lookup
Elementary particle wikipedia , lookup
Hydrogen atom wikipedia , lookup
Canonical quantization wikipedia , lookup
Wave function wikipedia , lookup
Particle in a box wikipedia , lookup
Renormalization group wikipedia , lookup
Path integral formulation wikipedia , lookup
Dirac equation wikipedia , lookup
Schrödinger equation wikipedia , lookup
Wave–particle duality wikipedia , lookup
Matter wave wikipedia , lookup
Relativistic quantum mechanics wikipedia , lookup
Theoretical and experimental justification for the Schrödinger equation wikipedia , lookup
1 Derivation of Schrödinger’s equation Mikhail Batanov, Associate Professor Department 207, Moscow Aviation Institute, Moscow, Russia Abstract In this article, a model of a material particle in chaotic motion (while maintaining a definite size and trajectory) is presented. On the basis of this model, the following is achieved: to express Planck’s constant ħ through the main features of a stationary random process; to justify the transition from the coordinate representation of the state of the particle to its momentum representation without invoking either the principles de Broglie waves or the Heisenberg uncertainty principle. to derive a form of the Schrödinger equation on the basis of the principle of extremum of the mean of the action of a particle in chaotic motion. The domain of the resulting equation is of course determined by its manner of derivation; we highlight the limits of the derived equation. An intermediate result, the determination of the density of the probability distribution of the nth derivative of an nth-order differentiable, random, stationary process can be applied in many areas of probability theory and statistical physics. Keywords: Schrödinger equation, Planck's constant, electron core, probability density distribution, derivative of a random process, particle, chaotic trajectory, coordinate representation. 2 1. Motivation and inspiration for this case of Schrödinger’s equation One of the most basic equations in modern physics is Schrödinger’s equation, best known under the form 2 2 i U ( x, y, z ) , t 2m (1) where = (x, y, z, t) is the wave function describing the state of an elementary particle, U (x, y, z) is the potential energy of the particle, is the reduced Planck's constant, and m is the mass of the particle. Schrödinger himself was putting together ideas of Louis de Broglie and Albert Einstein, looking for a generalization of the Hamilton-Jacobi equation. There have been many other interesting derivations of the equation since then. Although some of these derivations proceed out of classical theory, the end result is usually applied primarily to quantum theory. By examining Schrödinger’s original path to the equation, this paper investigates a derivation of an equation which we term a “generalized Schrödinger equation” which not only proceeds out of classical theory, but whose end result is also applicable to classical systems. The first historical steps that eventually led Schrödinger to his famous equation are well known. Louis de Broglie had noticed that that material particles in motion possess wave properties. In his doctoral thesis, Recherches sur la théorie des quanta, de Broglie compared the rectilinear trajectory of the free motion of a particle with a direct ray of light, and came to the conclusion that they are described by the same Hamilton-Jacobi equation, arising from the fundamental principle of "extremum of action". That is, the trajectory of the free motion of the particle and the beam of light are extrema for virtually the same functional, the action. This circumstance prompted Louis de Broglie to suggest that if the wave described by the equation w = expi(t – kr), (2) 3 (where is the angular frequency; k is the propagation vector; t is time; r is the dimensional vector), displays some properties of a particle, it is possible to reverse the assertion to say that a material particle in motion can correspond to a plane wave described by = expi(Et – pr)/, (3) where E is the kinetic energy of a moving particle, and p = mv is its momentum. In addition, Louis de Broglie was acquainted with experiments, carried out by his elder brother Maurice de Broglie, which were associated with the physics of X-ray radiation, as well as with the pioneering work of Max Planck and Albert Einstein on the quantum nature of radiation and absorption of light. This allowed him in 1923 to 1924 to propose that a moving particle can be associated with an oscillatory perturbation having frequency = E/ (4) = 2/|p|. (5) and with the wavelength This idea was supported by P. Langevin and Einstein, but most of the physics community reacted to it with skepticism. However, in the period from 1927 to 1930, several groups of experimenters (C. Davisson & L. Germer, and O. Stern, I. Estermann et al.) showed that the idea of the existence of matter waves, proposed by de Broglie, could be used to describe the phenomenon of the diffraction of electrons and atoms in crystals. In one of his early works of 1925 to 1926, Erwin Schrödinger, critical of the Bose-Einstein statistics formulation, wondered, "Why not start with the wave representation of the gas particles, and then impose on such ‘waves’ the quantization conditions ‘à la the Debye model’"? After that followed his central idea: "This implies none other than the need to take seriously into consideration the proposal of L. de Broglie and A. Einstein concerning the wave theory of moving particles." 4 The next article by Schrödinger already contained equation (1), marking the beginning of the intense development of quantum mechanics, together with the work of Max Planck, Albert Einstein, Niels Bohr, Werner Heisenberg and others. In this present paper, we backtrack a little in this history, going back to a model of a particle (having a given volume and continuous trajectory) in chaotic motion, leading to a new derivation of a form of the Schrödinger equation, with the result that the equation will now be applicable to classical systems. However, it is not only this equation which we can thus profit from, but also the means of derivation itself offers several advantages. To properly outline these advantages lies beyond the scope of this paper; we therefore leave this exposition for a future paper. 2. Model of a particle in motion Consider a particle occupying a small volume compared to that of its surrounding space. (Figure 1) By convention, we call this particle a "point". Suppose that this "point" is in constant chaotic motion around the reference point (serving as the origin of the coordinate system XYZ) under the influence of various mutually independent force factors. Examples of "points" in continuous chaotic motion may be atoms vibrating in a crystal lattice, flies fly- Fig. 1. The particle (“point”) in chaotic motion in the neighborhood of the central reference point so that its total mechanical energy E remains constant. ing in a jar, the nucleus vibrating inside a biological cell, a human embryo moving in the womb, tips of branches fluttering in a variable wind, and so forth. 5 In what follows, we suppose that such a chaotic motion of the "point" continues "forever" because its total mechanical energy E always remains constant: Е = T (x,y,z,t) + U (x,y,z,t) = const, (6) where T (x, y, z, t) represents the kinetic energy of the "point" due to its velocity, and U (x, y, z, t) is the potential energy of the "point" associated with a force tending to return the "point" to the central reference point (i.e., with an elastic force). Thus, in this model, each of the energies T (x, y, z, t) and U (x, y, z, t) of the "point" is a random function of time and its position relative to the "center". But these energies flow smoothly into each other so that their sum (i.e., the total mechanical energy E) always remains constant. If the speed of the "point" in chaotic motion in the vicinity of the central reference point (Figure 1) is low, then according to non-relativistic mechanics, it has kinetic energy T ( x, y, z, t ) p х2 ( x, y, z, t ) p 2у ( x, y, z, t ) p z2 ( x, y, z, t ) 2m . (7) For brevity, instead of (7) we write T (t ) p х2 (t ) p 2у (t ) p z2 (t ) 2m , (8) where pх(t), pу(t), pz(t) are the respective instantaneous values of the spatial components of the momentum of the "point" in chaotic motion, p p х2 p 2у p z2 , pi mvi m dxi m xi . dt (9) (10) The type of potential energy U (x, y, z, t) acting on the “point” is not specified. The action of the "point” S under consideration is defined in non-relativistic mechanics as follows (Landau & Lifshitz 1988) 6 t2 S (t ) [T ( p x , t ) U ( x, t )]dt Et . (11) t1 To simplify the calculations, we consider the one-dimensional case, without loss of generality. The three-dimensional case merely requires more integrations. Due to the complexity of the path of the “point" in motion, we are interested not in the action itself (11), but rather its average over time (resp., over its realizations). 1 S lim N N t2 N S (t ) [T ( p , t ) U ( x, t )dt E t. i i 1 x (12) t1 Recall that for an ergodic stochastic process, an average over time is equivalent to the average over its realizations. The plus sign in the integrand of (12) results from the fact that the mean potential energy is negative, as the “point” always tends to return to the reference point at the center of the sphere formed by the path of the point over time. (Figure 2). Finding the mean (12) is carried out over the realizations, taken for the same period of time Fig. 2. The total path of a particle in continuous chaotic motion as it tends towards spherical symmetry. t = t2 – t1 . The mean kinetic energy of a "point" in chaotic motion may be represented as 1 T ( px , t ) 2m ( p ) р dр x 2 x x , (13) where ρ(px) is the probability density function of the momentum component px of the material "points". 7 The mean potential energy of a "point" may be represented as U ( x, t ) ( x)U ( x)dx , (14) where ρ(х) represents the probability density function of the projection onto the x-axis of the “point” in chaotic motion around the fixed reference point. (Figures 1 and 2). Substituting (13) and (14) into the equation (12) for the mean of the action, we obtain t2 1 2 S ( p x ) р x dр x ( x)U ( x)dx dt E t. 2m t1 (15) For the further derivation of the Schrödinger equation, two auxiliary points are explained below. The first point, developed by the author (Gaukhman 2007 / 2008)), is dedicated to the definition of the distribution of the probability density of a stationary stochastic process which is n-times differentiable. The second item, the "coordinate representation of the average momentum of a particle" is borrowed from the work of D. Blokhintzev (1963) for ease of reference. 3. Determination of the probability density function of the n-th derivative of an n-times differentiable stationary stochastic process The key to the understanding of quantum mechanics and the limits of its application lies in the determination of the probability density function (PDF) of the derivative of a strictly stationary stochastic process, given that the PDF of the process itself is already known. This solution makes it possible to justify the quantum-mechanical transformations between a coordinate and a momentum representation without appealing to the concept of de Broglie waves. This is made possible due to the fact that the momentum of a material “point” is linearly related to the derivative of its coordinate px= m·x/t = mx׳. Moreover, the problem of determining the one-dimensional PDF ρ1[ξn(t)] , that is the n-th derivative of the n-times differentiable stationary stochastic process ξ (t), given only that its one-dimensional 8 PDF ρ1[ξ (t)] is known, arises in a series of problems in the fields of radio physics and statistical mechanics. We first consider the general properties of the first derivative of the stationary stochastic process ξ (t) by regarding its realizations (Figure 3). Fig. 3. Representation of the differentiable stationary stochastic process ξ(t). Figure 3 shows that the value of the random variable ξ (ti) at the interval ti is independent from the derivative (t i ) t1 of this process taken at the same value ti, ; that is, they are random variables t1 which are not correlated. This may be expressed analytically (Ritov 1976): 9 ti ti d 1 ti 2 1 d ti 2 0, dt 2 2 dt (16) where denotes taking the mean over the realizations. This result follows from the facts that the differentiation and averaging operators commute in this specific case, and that the average characteristics of the strictly stationary process, including its dispersion, are constant over time: t i const 2 Realizations of a stationary stochastic process ξ(ti), such as that shown in Figure 3, may be interpreted as the change over time of the projection onto the X-axis of the position of the “point” in motion (Figures 2 and 3). That is, x(t) = ξ(ti). Even in the case of the statistical independence of the random values ξi and ξI, there exists a connection between the PDF’s ρ1(ξi) and ρ1(ξi ). This results from the well-known procedure of obtaining the PDF of a derivative ρ1(ξi) for a given two-dimensional PDF of a stationary stochastic process (Ritov 1976, Tihanov 1982). 2 i , j 2 i , ti ; i , t j . (17) This is done by changing variables in (17): i k k ; j k k ; 2 2 ti t k ; 2 t j tk 2 , (18) where t j ti ; t k t j ti 2 with the Jacobian transform [J] = τ. Then, from the PDF (17) we obtain 2 k , k lim 2 k , t k ; k , t k . 0 2 2 2 2 (19) Further, integrating the resulting expression over ξk , we find the desired PDF of the derivative of the original process over the interval tk (Ritov 1976, Tihanov 1982): 10 1 ( k ) 2 ( k , k )d k . (20) The formal procedure given by (17) through (20) solves the problem of determining the PDF ρ1(ξ ) for a given two-dimensional PDF (17). However, a two-dimensional PDF is defined only for a very limited class of stochastic processes. It is therefore necessary to consider the possibility of obtaining a PDF ρ1(ξ ) for a given one-dimensional PDF ρ1(ξi). To solve this problem, we use the following properties of stochastic processes: 1] A two-dimensional PDF of any stochastic process can be represented as (Ritov 1976, Tihanov 1982) 2 i , ti ; j , t j 1 i , ti j , t j / i , ti , (21) where ρ(ξj, tj /ξi, ti) corresponds to the given PDF. 2] For the strictly stationary stochastic process, the following identity holds (Ritov 1976, Tihanov 1982) 1 i , ti 1 j , t j (22) 3] The given PDF of a stationary stochastic process as ti tends to tj degenerates into a delta function (Tihanov 1982) lim 2 i , t i / j , t j i j . 0 (23) Based on the above properties, we consider a stochastic process over the interval ] ti – τ0; ti + τ [ as τ tends to zero, using the following formal procedure. The PDF’s ρ1(ξi) = ρ1(ξi, ti) and ρ1(ξj) = ρ1(ξj, tj) can always be represented as the product of two functions: ( i ) i i 2 i ( j ) j j 2 j , where φ(ξi) represents the wave function of a random variable ξi over the interval ti. For a stationary stochastic process, we have the identity (24) 11 i j , (25) as is easily seen by taking the square root of both sides of (22). Then, according to (24), we obtain (25). Note that identity (25) is approximately true for the majority of non-stationary stochastic processes as τ tends to zero, that is, i , t i lim j , t j t i . 0 (26) When the condition (25) holds, equation (21) can be represented in the symmetric form i , j i j / i j , (27) where ρ(ξj /ξi) is the given PDF; in expanded form this becomes i , t i t k ; j , t j t k 2 2 i , t j t k j , t j t k / i , t i t k i , t i t k . 2 2 2 2 (28) Setting τ to zero in (28) in such a way that the interval τ contracts in equal measure on the left and the right to a point tk = (tj – ti)/2, then taking into account (23) and (27), we obtain lim ( i , j ) lim (i ) ( j / i ) ( j ) (ik ) ( ik ik ) ( jk ), 0 0 (29) where ξik is the result of the stochastic value ξ(ti) tending to the stochastic value ξ(tk) on the left, while ξjk is the result of the stochastic value ξ(tj) tending to the stochastic value ξ(tk) on the right. Integrating both sides of the expression (29) over ξik and ξjk,, we obtain ( ik ) ( jk ik ) ( jk )d ik d jk 1 . (30) Expression (30) is a formal mathematical identity out of the theory of generalized functions, taking into account the properties of the delta (δ)-function. In order to assign the expression (30) a physical content, one must specify the specific type of δ-function involved. 12 Therefore we now fix the form of a δ-function for a Markov stochastic process. Consider a continuous stochastic Markov process which satisfies the Einstein-Fokker equation (Tihanov 1982, Wentzel & Ovcharov 1991) ( j / i ) t B 2 ( j / i ) 2 , (31) where B is the diffusion coefficient. This parabolic differential equation has three solutions, one of which can be represented as (Tihanov 1982, Wentzel & Ovcharov 1991): 1 2 j , t j / i , ti 2 exp{iq( j i ) q 2 B(t j ti )}dq, (32) where q is the generalized parameter. As tj – ti = τ tends to zero, then from (32) we obtain one of the definitions of a δ-function 1 lim (i / j ) 0 2 exp{iq( jk ik )}dq ( j i ), (33) Since this result was obtained for the limiting case, that is, as τ tends to zero, then it is possible that the δ-function (33) can conform not only to Markov, but also to many other stationary stochastic processes. Substituting the δ-function thus obtained (33) into (30) yields 1 (ik ) ( jk ) 2 exp{iq( jk ik )}dq d ik d jk . (34) Changing the order of integration in (34), we obtain 1 2 ( ik ) exp{iq ik }d ik 1 2 ( jk ) exp{iq jk )d jk dq . (35) Take into account that, according to (25), φ(ξik) = φ(ξjk), and also the properties of δ-functions, 13 ξik = ξjk = ξk. Under these conditions, the expression (35) takes the form (q) (q)dq 1, (36) where q 1 2 1 q 2 * ( ) exp{iq }d , k k k (37) ( k ) exp{iq k }d k . (38) The integrand (q) (q) in (36) satisfies all the requirements of the PDF ρ(q) of the random vari- able q: (q) (q) (q) (q) . 2 (39) We now investigate the random variable q. We first reconsider (32). The result of the integration on the right side of this expression does not depend on the variable q. Therefore it may be considered as a sort of generalized frequency. However, both the physical statement of the problem as well as the mathematical formalism in the expression (32) impose on q the following restrictions: 1) The variable q must characterize a stochastic process in the interval under investigation ] ti – τ; ti + τ [ as τ tends to zero; 2) The variable q, according to the mathematical notation on the right side of (32), must belong to the set of real numbers (q R), having the cardinality of the continuum; that is, it must have the possibility to take any value in the range ]– ∞, ∞ [; 3) q must be stochastic 14 All these requirements are satisfied by any of the following random variables associated with a stochastic process in the interval under investigation τ: i k 2 k n i , i , . . . , n i . t t t n (40) But these random variables do not equally characterize the process. We consider one of the realizations of the test process. The function ξ(t) (Figure 3) in the interval ]ti ; tj = ti + τ[ for τ < τcor (where τcor is the correlation radius of a stochastic process) may be expanded as a Maclaurin series: t j ti (ti ) (ti ) 2 2 ... (n) n! n ... (41) We rewrite the expression (41) in the form j i i i 2! ... ( n ) n 1 n! ... (42) As in (33), we let tend to zero, whereby (42) reduces to the identity lim j i 0 k (43) In this way, the only random variable satisfying all the above-mentioned requirements in the interval under investigation ]ti = tk – τ/2; tj = tk + τ/2[ as τ tends to zero is the first derivative of the original stochastic process ξ'k over the interval tk. Therefore we may assume that the random variable q in (32) through (39) is directly proportional to ξ'k, that is q k , (44) where 1/η is the proportionality coefficient. Substituting (44) into (36) through (39), we obtain the following procedure as required for determining the PDF of the derivative ρ(ξk) of a stationary Markov process ξ(t) over the interval tk , given the 15 one-dimensional PDF ρ(ξk) over the same interval. (This procedure may be valid for other stochastic processes as well.) 1] Express the given one-dimensional probability distribution ρ(ξ) as the product of two wave functions φ(ξ): ( ) . (45) 2] Two Fourier transforms are then carried out: ( ) 1 2 ( ) ( ) exp{i /}d , (46) 1 2 ( ) exp{i / }d . (47) 3] Finally, for any given interval of a stationary Markov process we get the desired derivative of the PDF, as follows: ( ) * . 2 (48) As we have already remarked, the procedure given by (45) through (48) can be applied not only to the stationary Markov processes, but to many other stationary processed for which the δ-function in (30) takes the form (33) as τ tends to zero. To find the physical interpretation of the proportionality factor 1/η, we make a comparison to known results. This method is not mathematically perfect, but permits us to quite efficiently obtain a sound result of practical importance. We consider a stationary Gaussian stochastic process. Over every interval of the process, the random variable ξ obeys the Gaussian distribution: 16 ( ) 1 2 2 exp ( a ) 2 / 2 2 , (49) where ξ2 and аξ are the variance and expected value, respectively, of the process ξ(t). Subjecting the PDF (49) to the sequence of operations (45) through (48), we obtain its derivative: ( ) ( ) 1 2 / 2 1 2 / 2 2 2 exp 2 / 2 2, (50) 2 2 exp 2 / 2 2, (50) On the other hand, with the help of a well-known procedure indicated in (17) through (20) for an analogous case, we obtain (Ritov 1976, Tihanov 1982) ( ) 1 2 2 exp 2 / 2 2 , (51) where σξ’ = σξ /τcor, and τcor is the correlation radius of the initial stochastic process ξ(t). Comparing the expressions (50) and (51) we find that 2 2 cor (52) These PDF’s are in complete agreement with one another. Expression (52) is obtained for a Gaussian stochastic process, with σξ being the standard deviation and τcor being the correlation radius, the basic characteristics of a stationary stochastic process. The rest of the initial and central moments in the case of a non-Gaussian distribution with the random variable ξ (t) will give a small but insignificant contribution to the expression (52), so it can be asserted with a high degree of confidence that this expression is applicable to a wide class of stationary stochastic processes. 17 It should be noted that in statistical physics and quantum mechanics, the transition from the coordinate representation of a function of an elementary particle state to its momentum representation is effected by a formal process almost completely analogous to the procedure (45) through (48). The difference is only in determining the proportionality factor 1/η. In quantum mechanics it is well known that if the projection onto the x-axis of the position of a free elementary particle (e.g., an electron) is described by a Gaussian distribution (Landau & Lifshitz 1989) ( х) х 1 2 2 х2 2 x exp 2, 2 x (53) where σx is the standard deviation of the projections of the positions of an elementary particle onto the x-axis in the neighborhood of the mean (that is, the “center” of the system). Then, as the result of operations analogous to that of (45) through (48), it turns out that the PDF of the momentum component рх of an elementary particle is also Gaussian (Landau & Lifshitz 1989) ( px ) px 1 2 2 2px p x2 exp 2 2 px (54) with the standard deviation px 2 x , (55) where the reduced Planck constant = 1.055·10-34 J/Hz. If we now take into account that the momentum component of an elementary particle (e.g. an electron) px is equal to px me dx me x , dt (56) 18 where me is the electron rest mass, then taking into account (55), the PDF (54) becomes ( x) 1 2 /(me 2 x )2 x 2 exp 2. 2 /(me 2 x ) (57) Comparing (50) to (57), while taking into consideration (52) and that ξ' = х', we find that for the given case 2 2 ex ex , me (58) where ex 2 2me ex 2 0,91 10 30 2 2 ex 1,73 10 4 ex 1,055 10 34 (59) is the correlation radius of a stationary stochastic process, which is the result of the projection of the stochastic motion of the "point" (e.g., an electron) onto the x-axis near the stationary "center" of the system (Figures 1 and 2). σex is the standard deviation of the projection onto the x-axis of the “point” (e.g., an electron) in chaotic motion around the mean (that is, the relative “center” of the system). From (58) it follows that the Planck constant is not a fundamental physical constant, but rather a random variable which may be expressed in terms of more basic parameters of a stationary stochastic process. 2 2par, x m par, x , (60) where for a general case: σpar,x is the standard deviation of the projection onto the x-axis of a particle ("point") in chaotic motion in the neighborhood of the point of reference (i.e., the “center” of the system). τpar,x is the correlation radius of a given stationary stochastic process. 19 For a wide range of applications, the expression (60) is in itself very important, as is the related ratio (52), which in the general case may conveniently be represented as follows: ч 2 2par, x par, x m (61) Note the following interim conclusions: 1] The quantum-mechanical transition from the coordinate representation to the momentum one is not only applicable to the processes in the world of elementary particles, but also to any stationary Markov stochastic processes (and probably many other stochastic processes), both in the microcosm and in the macrocosm. For example, a tree branch, constantly moving chaotically around its middle position (the point of reference serving as the "center") by the rapidly changing direction of wind gusts, behaves similarly to elementary particles in the "potential well". The fluctuations of these movements of the branch would also have a discrete (quantum) average set of states, depending on the intensity of the wind gusts. With weak wind gusts, the branch generally fluctuates near the central reference point, in a way that the position of its tip can be described by a Gaussian distribution. Wind gusts of greater intensity will see the tip of the branch rotate in a circular motion; at still higher intensities, it would essentially describe a figure eight, and so forth. Depending on the strength of the wind, the tip of the branch may describe a discrete set of Lissajous figures. In other words, the quantum-mechanical formalism is not an exclusive feature of the microcosm; it is also applicable to the statistical description of many stochastic processes in the macrocosm. 2] The algorithm described by (45) through (48) of the transition from the coordinate representation ρ(ξi) to the momentum representation ρ(mξi) and vice versa was obtained in the context of a particular δ-function (33). It would be interesting to analyze what would be the result in the case of other types of δ-function. 20 3] On the basis of the foregoing, we may obtain the PDF ρ(ξi''), which is the second derivative of a differentiable stochastic process which is differentiable to at least order two. In this case, we should consider not the process (t) itself, but its first derivative (t) = (t)/t. Then the distribution of the second derivative may be obtained according to the same procedure, (45) through (48), whereby ρ(ξi).would replace ρ(ξi) in those equations. Analogously, we also may obtain the PDF ρ(ξi(n)) of any derivative of n-times differentiable stationary stochastic process with the help of the following recursive procedure: ( ( n1) ) n1 n1 ; n 2 n * 1 ( ( n1) 1 2 i ( n) ( n1) ( n1) ) exp d ; n (62) i ( n) ( n1) ( n1) d ; n ( ( n1) ) exp (63) (64) ( ( n) ) n * n , where n 2 2( n 1) cor ( n 1) (65) where ( n 1) , cor n 1 are the variance and the correlation radius, respectively, of the given n–1 times 2 differentiable stationary stochastic process. 4] The procedure (45) through (48), fully analogous to that of the quantum-mechanical transition from the coordinate representation of a quantum system to its momentum representation, is obtained here on the basis of the study of realizations of a common stationary stochastic process, that is, without the use of phenomenological principles of wave-particle duality. 21 5] In the aspects of the diffraction of particles in a crystal lattice that we will be covered by the version of the Schrödinger equation that we will develop below, we can circumvent the use of deBroglie waves. One may study paragraph 2.9.6 of (Gaukhman 2008), where one will find the derivation of the formula to create a three-dimensional diagram (indicatrix) of the scattering of particles on the surface of a multilayer periodic crystal: ( , / , ) 4 n12 k к sin 2 [ n1 / 2 k к (a 2 b 2 ) / c 2 / 2] ca b a b c ba ab [( n1 ) 2 k к2 (a 2 b 2 ) / c 2 ]2 c2 a2 b2 , (66) where a cos cos cos cos , b cos sin cos sin , c sin sin , av sin cos , bv sin v sin , cv cos v , a cos v sin , b cos v cos Angles , , ω and ν are shown in Figure 4; kк = rcor n11/2/ (0,066 l1 ), where: l1 signifies the thickness of one layer, i.e. one sinusoidal equipotential surface; n1 equals the number of layers effectively involved in the scattering of the particles; rcor is the average radius of curvature of a sinusoidal equipotential surface. For a single crystal, all the sinusoidal equipotential surfaces have the same, that is, rcor. Therefore in this case rcor signifies the effective cross section of the scattering of electrons by the atoms of a crystal. The results of a calculation using (66) at an angle of incidence of particles on the surface of the crystal using the values = 45о, azimuth angle = 0о, rcor = 0.0000000001 = 10–10 cm, l1 = 0.000000001 = 10–9 cm, n1 = 1940 (layers) are shown in Figure 4: 22 Fig. 4. Diagram (indicatrix) of particle (electron) scattering through 1940 layers of sinusoidal equipotential surfaces of a crystal, calculated according to (66) using MathCad14 (Gaukhman 2008, Batanov 1994). 4.The coordinate representation of the average particle momentum The following calculations are gathered here, following (Blokhintsev 1963), in order to make this paper as self-contained as possible. Let us first recall the properties of the Dirichlet integral appearing in the theory of Fourier integrals and the theory of generalized functions (Blokhintsev 1963): lim 1 b ( z) k a 0, if a, b 0 or a, b 0, sin kz dz z 0, if a 0, b 0 . (67) whereby 1 sin kz z k z lim (68) This is one of many forms of a δ-function. We now consider the calculations for reducing dimensions by demonstrating the equality (Blokhintsev 1963): p xn ( p x ) p zn dpx ( p x ) p zn ( p x )dpx ( x, t ) i ( x, t )dx, x n (69) 23 where the exponent n is a positive integer. Taking the mean over time (or realizations) to the power of n momentum components рхn = (m·x/t)n = (mx)׳n; (70) where x and ( p ) are the wave functions which were introduced in (24). х [ x x ] and (48) [ ( px ) px mx ], and, following (46) and (47), are related to one other (assuming a stationary stochastic process) via a Fourier transform: i p x mx ( х) xx i ч px x е е dx ( х) dx ; 1/ 2 1/ 2 (2 ) ( 2 ) i xx ч i i (71) px x е е * p x mx ( х) dx ( х ) dx ; (2 )1/ 2 (2 )1/ 2 x ( x) i xx ч px x е е dx ( р х ) dр х ; 1/ 2 1/ 2 (2 ) ( 2 ) i xx ч i (72) px x е е * x ( x) d x ( р ) х (2 )1/ 2 dрх , (2 )1/ 2 where the parameter ηpar is defined by (61) par 2 2par, x par, x . m (73) In order to prove (69), for ψ(рx) and ψ*(рx) we substitute their respective expressions in terms of integrals (71) (Blokhintsev 1963): p x p x i x k i x l e e n n px ( xk ) dxk px ( xl ) dxl dpx . 1/ 2 1/ 2 ( 2 ) ( 2 ) An immediate consequence of this is that (74) 24 n x p e i p x xl i xl n i px xl e . (75) Substituting (75) into (74) we obtain: 1 p 2 n x p x i x k ( xk )e dxk ( xl ) i xl n i px xl e dxl dpx . (76) We perform an integration by parts n times, starting from the second integral in the integrand. In doing so, we assume that ψ (x) and its derivatives vanish at the boundaries of integration, that is, at x = ± ∞. Following these steps, we find (Blokhintsev 1963): 1 p 2 n x p x p x i x k i x l ( xk )e dxk e i xl n ( xl )dxl dpx (77) Changing the order of integration, we first integrate over px (Blokhintsev 1963): 1 p ( x ) dx k k i xl 2 n x n p ( x x ) i x k l ( x j )dxl e dpx . (78) We introduce the variables ξ = рx /h, z = xk – xl. In the last integral in (78), we perform the integration over ξ between the limits – k to + k; then passing to the limit k → ∞, this expression takes the form (Blokhintsev 1963): n sin kz p i ( x) dx lim ( x z )dz k x z n i ( x)dx ( x z ) ( z )dz. x n x Based on the properties of the Dirichlet integral (67), when a = –∞; b = + ∞, and ψ(z) = ψ*(x + z), we have (Blokhintsev 1963): (79) 25 n n p i ( x) ( x)dx ( x) i ( x)dx x x n x (80) or p xn ( x, t ) i ( x, t )dx , x n (81) where ( x, t ) ( x) expiu (82) ( x, t ) ( x) exp iu (u is an arbitrary real number); thereby the expression (69) is proven (Blokhintsev 1963). Using (70) and (73) from the expression (81) we obtain n n dx x ( x ) i par ( x )dx . x dt n (83) The generalization to three dimensions then increases the number of integrations (Blokhintsev 1963). 5. Derivation of Schrödinger’s equation We return to the average action of the particle ("point") in chaotic motion (15) 1 2 S ( p ) р dр ( x ) U ( x ) dx dt E t , x x x 2m t1 t2 (84) Imagine the action (84) in coordinate form. To do this, perform the following steps. 1] We write the PDF ρ(х) as a product of two wave functions using ψ(х,t): ( x) x, t * x, t , Et , ( x, t ) ( x) expi (85) (86) 26 Et ( x, t ) ( x) exp i . (87) 2] We use the coordinate representation of the mean momentum, raised to the power n (81). Thus in particular we have p x2 2 ( p x ) p z dpx ( x, t ) i ( x, t )dx , x 2 (88) 3] Using (88), we represent the mean kinetic energy of the "point" (13) in the form 1 2 1 2 2 ( х, t ) 2 T рx ( p x ) р x dр x ( x, t ) dx, 2m 2m 2m x 2 (89) 4] By taking (85) into consideration, we can now represent the mean potential energy of the "points" (14) in the form U ( x, t )U ( x) ( x, t )dx, (90) 5] It is then easily seen that EЕ i ( x)e iEt / ( x)e iEt / dx const, t (91) or, taking into consideration also (86) and (87) E Е ( x, t ) ( x, t ) dx t (92) 6] Substituting expression (89), (90) and (92) into (84), we may express the mean of the positions of the particles ("point") in chaotic motion in coordinate form 2 2 ( x, t ) ( x, t ) S ( x, t ) dx ( x, t )U ( x) ( x, t )dx i ( x, t ) dxdt 2 2 m x t t1 t2 (93) 27 or t2 S 2 2 ( x, t ) ( x, t ) ( x , t ) ( x , t ) U ( x ) ( x , t ) ( x , t ) i 2 2m dxdt t x t1 (94) Finding an extremum for the mean of the action (94) requires setting the first variation to zero t2 S 2 2 ( x, t ) ( x, t ) ( x , t ) ( x , t ) U ( x ) ( x , t ) ( x , t ) i 2 2m dxdt 0 t x t1 (95) The extremum of the functional (95), that is, the function (x, t), for which the mean of the action (95) takes an extreme value, is determined by the Euler-Poisson equation (Èl'sgol'ts 1969). For the Lagrangian L, which is the integrand in the action functional z z 2 z 2 z 2 z dxdt , where z = (x, t), S L x, t , z, , , 2 , 2 , x t t x tx (96) the equation takes on the form Lz 2 Lp L g 2 L x t x r 2 2 Ls 0 , L t xt t 2 (97) where z z 2z 2z p ; g ; r ; s , x t tx x 2 z p g L p L px L pz L pp Lgp x x x x complete the partial derivative with respect to x. Analogously, (98) 28 L p g Lg L gt L gz L gp L gg t t t t (99) and so forth. Using the integrand of the mean of the action (95) L 2 2 ( x, t ) ( x, t ) ( x, t ) ( x, t )U ( x) ( x, t ) ( x, t )i 2 2m t x we define Lz 2 2 x x 2 x U x i ; 2 2m x t L p 0; x Lg 2i x ; x t 2 Lt 0; t 2 2 Ls 0; xt 2 h 2 2 x L . r x 2 2m x 2 Substituting these expressions into (97), we obtain the desired equation for determining extrema ψ (х,t) of the mean of the action functional (95) i ( x, t ) 2 2 ( x, t ) + U ( x) ( x, t ) , t 2m x 2 (100) where (x,t) (x,t) = | (x)| 2 = (x) is the PDF of the projections onto the x-axis of the position of the given particles ("points") in chaotic motion around the point of reference serving as "center" of the system in such a way that its total mechanical energy E is always constant (E = const), while the projection x(t) itself is a stationary stochastic process. The generalization to three dimensions only increases the number of integrations; thus we have i where (r , t ) 2 2 (r , t ) + U (r , t ) (r , t ) , 2 t 2m r (101) 29 (r , t ) (r ) exp i Et , where r is the radius vector starting at the above mentioned "center" of the configuration (r2 = x2 + y2 + z2) (Figure 1). Equation (101) is none other than the Schrödinger equation (1) with the Born understanding of the meaning of the wave function. But in this case, "the Planck constant” ħ is not a fundamental constant, but rather the parameter which can be expressed in terms of the ratio of the average characteristics of the stationary stochastic process being investigated, as above in (60) 2 2par, x m par, x , Dividing both sides of the equation (101) by ħ we get (r , t ) 2 (r , t ) i + U (r , t ) (r , t ). 2 t 2m r Taking (61) into consideration, this equation takes the form par 2 ( r , t ) ( r , t ) i + U (r , t ) (r , t ) , 2 t 2 r (102) where par 2 2par,r par,r , (103) whereby par,r 1 2par, x 2par, y 2par, z 3 (104) is the mean standard deviation of the positions of particles (material "points") in chaotic motion around the point of reference acting as "center" (Figure 1), while par , r 1 ( par , x par , y par , z ) 3 (105) 30 is the mean correlation (or rather autocorrelation) radius of this stochastic process. We term Equation (102) a “generalized Schrödinger equation”, as it is suitable to describe the most probable average states of point objects in chaotic motion, on microcosmic as well as on the macrocosmic scales, for a given stationary stochastic process with its total mechanical energy held constant. Equation (102) can be used to describe or approximate phenomena which are considered primarily classical, such as average behavior of an electron in a potential well, the nucleus in the cytoplasm of biological cells, the center of the embryonic mass in the womb, the core in the bowels of the planet, flies trapped in a jar, and so on. Each of these stable stochastic processes has the ability to move from one steady state to another with the absorption or release of a particular portion of its total mechanical energy. It would be an interesting endeavor to construct the scaling up of such applications of (102) in a fractal manner, but this is beyond the scope of this paper. The interest in this approach lies not only in the derivation of Schrödinger’s equation, but also that one could also use this approach for derivations of other key equations such as the Klein-Gordon equation, the Dirac equation, and so forth. The algorithms for these derivations follow steps which are analogous to the ones given in this article: 1) express the instantaneous deterministic action of the particle; 2) find the mean of this action, 3) represent all averaged terms in the integrand of the mean action by their respective probability density functions ρ(х) and/or ρ(px); 4) switch all terms of the Lagrangian (the mean of the action) to a coordinate or a momentum representation; 5) determine the equation for the extrema of the resulting functionality (the mean of the action) through methods of the calculus of variations. 31 The significance of this derivation of the generalized Schrödinger equation (102) is, in the author’s opinion, as follows: The derivation highlights the limits and conditions of the application of this equation There is no need to apply either Heisenberg’s uncertainty principle or de Broglie waves, since, for the derivation of equation (102), the procedure (45) through (48) is used, and this procedure is completely analogous to the transition from the coordinate representation to the momentum one, and vice versa. However, this procedure is obtained based only on the analysis of the properties of a stationary stochastic process, without the involvement of the above hypotheses. The quantity ħ/m (“Planck's constant” divided by the mass) is obtained through the definitions of variance and of correlation coefficient (i.e. through their respective mean characteristics) while investigating stationary stochastic processes. Therefore, the generalized Schrödinger equation (102) does not contain an explicit term for the particle’s mass. The advantages of this fact are outlined in Paragraph 1.7.10 in (Gaukhman 2007) and Section 7 in (Gaukhman 2008). Mass as a derived rather than a fundamental concept is of course not new. Recall that in quantum field theory, it is possible to determine, more or less correctly, the rest mass of a particle by a structure, altogether not simple, of pseudosymmetry breaking involving Higgs bosons. There are a number of formulas used almost exclusively in the quantum realm. However, this does not mean that these same equations cannot be extended to the classical realm, albeit with a correspondingly altered interpretation. In order to make the case for a classical equivalent of such an equation, it is more convincing when the equation in question may be derived purely in the classical domain, without help from concepts which are absent from the classical world. In this paper we have done just that, keeping to the classical ideas of a distinct trajectory and size of a particle, ideas which are no longer 32 valid in the domain covered by quantum physics. Historically, the reasons for abandoning the classical assumptions were many; among them: The instability of the distributed (non-point) charge: if the charge had some kind of size, then between its near points there would be such a huge electrostatic repulsion force that the charge with nonzero volume would immediately fly apart into small pieces, in which even greater repulsive forces would act, and so forth. The indivisibility of elementary particles (earlier derived from the reasoning that since, if such a particle had a distinct size, it would be possible to cause an impact onto one of its parts in such a way that this part would separate from the rest, but this does not happen). Also, elementary particles would then be distributed over the internal characteristics of the space, such as stress, strain, temperature, etc. But there is no evidence of any of this in quantum physics. It gives elementary particles (as discrete particles, i.e. mathematical points) only those discrete characteristics as: spin, charge, magnetic moment, and so forth. The global nonlocal (wave) nature of elementary particles, as in some experiments in which they generally manifest themselves as "spread" throughout space. These problems were solved by abandoning the concepts of a definite path and position for the quantum world. These solutions then were presumed to separate the quantum and the classical realms when applying certain equations. However, we maintain that there are possibilities to keep those concepts for the classical case, yet still apply the same equations. We do this by showing that the above apparent dichotomy between the quantum and the classical cases could have been avoided. To do this, we show how the classical assumptions and the above apparent divergences from the classical theory are actually not incompatible with one another for the equations (as equations) in question, even though the interpretation may differ. There are undoubtedly various ways to do this; one approach is to present the 33 solution to the above difficulties in a framework of a fully geometrized theory. The derivation presented in this paper is a detail from one such theory (termed the "Algebra of signatures") which is as of yet only presented in Russian at (www.alsignat.narod.ru ). The author hopes to continue unfolding this theory in a series of papers in English. Acknowledgements The main thesis of this article was first published in 1990 in part in series of discussions which appeared in (Batanov 1990 / 1994). For this I thank my mentors, Drs. A. Kuznetsov and A.I. Kozlov. I also wish to thank Drs. A.A. Rukhadze and A.M. Ignatov for their decision to publish one of the versions of this article in the journal "Engineering Physics" № 3, 2016. I also thank David Reid for his creative and highly professional approach to the present translation and editing of this article into its present form. Bibliography [1] Batanov, M.S. (1990) Kachestvenno novoye ponimaniye strukturnoy organizatsii materii (vyvod uravneniya Shredingera). [A qualitatively new understanding of the structural organization of matter (Derivation of the Schrödinger equation)] in Problemy tekhnicheskoy ekspluatatsii i sovershenstvovaniya REO.[Technical operation problems and improvement in REO] Moscow, USSR: Mezhvuzovskiy sb. nauch. tr. RIO MIIGA. pp 145 - 156. [In Russian] [2] Batanov, M.S. (1994) Vliyaniye podstilayushchey poverkhnosti na tochnostnyye kharakteristiki kvazidoplerovskogo pelengatora [The influence of land surface on precision characteristics of a quasiDoppler direction finder.] Doctoral Dissertation. Moscow, USSR: Moscow State Technical University. [In Russian] [3] Blokhintsev, D.I. (1963) Osnovy kvantovoy mekhaniki [Principles of Quantum Mechanics]. Moscow, USSR: Vysshaya shkola. [In Russian] 34 [4] Èl'sgol'ts, L.E. (1969) Differentsial'nyye uravneniya i variatsionnoye ischisleniye [Differential equations and calculus of variations]. Moscow, USSR: Nauka. [In Russian] [5] Gaukhman, M. Kh. (2007) “Pustota” (zheltaya Alsigna). ["Void" (yellow Alsigna)]. In Algebra signatur [Algebra of signatures] Moscow, Russia: URSS. p. 308. (available in www.alsignat.narod.ru.) [In Russian] [6] Gaukhman, M.Kh. (2008) “Chastitsy” (zelennaya Alsigna). ["Particles" (green Alsigna)]. In Algebra signatur [Algebra of signatures]. Moscow, Russia: Librokom. (available in www.alsignat.narod.ru.) [In Russian] [7] Gaukhman, M.Kh. (2009) “Gravitatsiya” (golubaya Alsigna). ["Gravity" (blue Alsigna). In Algebra signatur [Algebra of signatures] Moscow, Russia: Librokom. (available in www.alsignat.narod.ru.) [In Russian] [8] Batanov, M.S. (2010) Algebra signatur [Algebra signature] in www.alsignat.narod.ru , Nicecorp.ru. Web. [In Russian] [9] Landau, L.D. & Lifshitz, E.M. (1988) Mekhanika [Mechanics]. Moscow, USSR: Nauka. [In Russian] [10] Landau, L.D.& Lifshitz, E.M. (1989) Kvantovaya mekhanika. Nerelyativistskaya teoriya [Quantum mechanics. Non-relativistic theory]. Moscow, USSR: Nauka. [In Russian] [11] Prigogine, I. & Stengers I. (2000) Poryadok iz khaosa [Order out of chaos]. Moscow, Russia: Editorial URSS. [In Russian] [12] Prigogine, I. & Stengers, I. (2001) Vremya, khaos, kvant [Time, chaos, quantum]. Moscow, Russia: Editorial URSS. [In Russian] [13] Ritov, S.M. (1976) Vvedeniye v statisticheskuyu radiofiziku, Ch. 1 [Introduction to Statistical Radio Physics, Part 1]. Moscow, USSR: Nauka. [In Russian] 35 [14] Tihanov, V.I. (1982) Statisticheskaya radiofizika [Statistical Radio Physics]. Moscow, USSR: Radio i svyaz', [In Russian] [15] Wentzel, E.S. & Ovcharov, L.A. (1991) Teoriya sluchaynykh protsessov i yeye inzhenernyye prilozheniya [The theory of random processes and its engineering applications]. Moscow, USSR: Nauka. [In Russian]