Creation of high-energy phonons by four-phonon processes in anisotropic phonon systems of He II

The problem of the creation of high-energy phonons (h-phonons) by a pulse of low-energy phonons
 (l-phonons) moving from a heater to a detector in superfluid helium, is solved. The rate of h-phonon creation
 is obtained and it is shown that created h-phonons occupy a much smaller sol...

Повний опис

Збережено в:
Бібліографічні деталі
Опубліковано в: :Физика низких температур
Дата:2007
Автори: Adamenko, I.N., Kitsenko, Yu.A., Nemchenko, K.E., Slipko, V.A., Wyatt, A.F.G.
Формат: Стаття
Мова:Англійська
Опубліковано: Фізико-технічний інститут низьких температур ім. Б.І. Вєркіна НАН України 2007
Теми:
Онлайн доступ:https://nasplib.isofts.kiev.ua/handle/123456789/127809
Теги: Додати тег
Немає тегів, Будьте першим, хто поставить тег для цього запису!
Назва журналу:Digital Library of Periodicals of National Academy of Sciences of Ukraine
Цитувати:Creation of high-energy phonons by four-phonon processes in anisotropic phonon systems of He II / I.N. Adamenko, Yu.A. Kitsenko, K.E. Nemchenko, V.A. Slipko, A.F.G. Wyatt // Физика низких температур. — 2007. — Т. 33, № 5. — С. 523-537. — Бібліогр.: 22 назв. — англ.

Репозитарії

Digital Library of Periodicals of National Academy of Sciences of Ukraine
_version_ 1860241800284864512
author Adamenko, I.N.
Kitsenko, Yu.A.
Nemchenko, K.E.
Slipko, V.A.
Wyatt, A.F.G.
author_facet Adamenko, I.N.
Kitsenko, Yu.A.
Nemchenko, K.E.
Slipko, V.A.
Wyatt, A.F.G.
citation_txt Creation of high-energy phonons by four-phonon processes in anisotropic phonon systems of He II / I.N. Adamenko, Yu.A. Kitsenko, K.E. Nemchenko, V.A. Slipko, A.F.G. Wyatt // Физика низких температур. — 2007. — Т. 33, № 5. — С. 523-537. — Бібліогр.: 22 назв. — англ.
collection DSpace DC
container_title Физика низких температур
description The problem of the creation of high-energy phonons (h-phonons) by a pulse of low-energy phonons
 (l-phonons) moving from a heater to a detector in superfluid helium, is solved. The rate of h-phonon creation
 is obtained and it is shown that created h-phonons occupy a much smaller solid angle in momentum space,
 than the l-phonons. An analytical expression for the creation rate of h-phonon, along the symmetry axis of a
 pulse, are derived. It allows us to get useful approximate analytical expressions for the creation rate of
 h-phonons. The time dependences of the parameters which describe the l-phonon pulse are obtained. This
 shows that half of the initial energy of l-phonon pulse can be transferred into h-phonons. The results of the
 calculations are compared with experimental data and we show that this theory explains a number of experimental
 results. The value of the momentum, which separates the l- and h-phonon subsystems, is found.
first_indexed 2025-12-07T18:30:39Z
format Article
fulltext Fizika Nizkikh Temperatur, 2007, v. 33, No. 5, p. 523–537 Creation of high-energy phonons by four-phonon processes in anisotropic phonon systems of He II I.N. Adamenko1, Yu.A. Kitsenko2, K.E. Nemchenko1, V.A. Slipko1, and A.F.G. Wyatt3 1 Karazin Kharkov National University, 4, Svobody Sq., Kharkov 61077, Ukraine 2 Akhiezer Institute for Theoretical Physics National Science Center «Kharkov Institute of Physics and Technology» National Academy of Sciences of Ukraine 1 Academicheskaya Str., Kharkov 61108, Ukraine 3 School of Physics, University of Exeter, Exeter EX4 4QL, UK E-mail: a.f.g.wyatt@exeter.ac.uk Received July 31, 2006 The problem of the creation of high-energy phonons (h-phonons) by a pulse of low-energy phonons (l-phonons) moving from a heater to a detector in superfluid helium, is solved. The rate of h-phonon creation is obtained and it is shown that created h-phonons occupy a much smaller solid angle in momentum space, than the l-phonons. An analytical expression for the creation rate of h-phonon, along the symmetry axis of a pulse, are derived. It allows us to get useful approximate analytical expressions for the creation rate of h-phonons. The time dependences of the parameters which describe the l-phonon pulse are obtained. This shows that half of the initial energy of l-phonon pulse can be transferred into h-phonons. The results of the calculations are compared with experimental data and we show that this theory explains a number of experi- mental results. The value of the momentum, which separates the l- and h-phonon subsystems, is found. PACS: 67.70.+n Films (including physical adsorption); 68.08.–p Liquid–solid interfaces; 62.60.+v Acoustical properties of liquids. Keywords: phonon–phonon interaction, liquid helium, anisotropic phonon system. 1. Introduction The dispersion relation, which has an important role for the processes of phonon-phonon interactions in superfluid helium (He II), can be written as � �� �� �cp p( ),1 (1) where c is sound velocity, � and p correspond to phonon energy and momentum, � �( p is a function which de- scribes the deviation of the spectrum from linearity which is small (| ( |� �p �� 1), but it nevertheless completely deter- mines the possible interactions of phonons. When p pc� , where ~pc �10 K (here and below ~ /p cp k B� ) at zero pressure, the function � 0 and the dispersion is anomalous. Then the conservation laws of energy and momentum allow processes which do not con- serve the number of phonons. The three-phonon process (3pp) is fastest of these processes and involves one phonon decaying into two or two interacting phonons combining into one. When p pc the function � � 0 and the dispersion is normal. Then the decay processes are prohibited by the conservation laws of energy and momentum, and the fast- est scattering is by four-phonon processes (4pp or 2 2 ). The three-phonon rate �3pp is much higher than the four-phonon rate �4 pp � �3 4pp pp (2) and this inequality of rates causes the phonons in superfluid helium to separate into two subsystems: a sub- system of high-energy phonons (h-phonons) with p pc in which equilibrium is attained relatively slowly, and a subsystem of low-energy phonons (l-phonons) with p pc� in which the equilibrium occurs relatively quickly. On the time scale of the concerned problem, the equilib- rium in the subsystem of l-phonons occurs instantly and the © I.N. Adamenko, Yu.A. Kitsenko, K.E. Nemchenko, V.A. Slipko, and A.F.G. Wyatt, 2007 phonon distribution of this anisotropic phonon system is given by the Bose–Einstein distribution function with two parameters, temperature and drift velocity [1,2]. The possibility of two-phonon subsystems with differ- ent relaxation times leads to interesting phenomena in anisotropic phonon systems of He II. Such systems are created by a heater immersed in the liquid helium, which is at such a low temperature that the thermal excitations can be neglected. When a short current pulse, of duration t p , is given to the heater, a phonon pulse in the liquid is formed by fast three-phonon processes. This l-phonon pulse moves in a «phonon vacuum», from the heater to the detector, with a velocity very near to c [3]. In momentum space, the phonons in the pulse are mainly located in a narrow cone with a solid angle � p �� 1. The transverse dimensions of this phonon pulse, near the heater, are about dimensions of the heater (typi- cally 1 1 mm), and the longitudinal dimension, which is determined by the duration of current pulse, is ct p : typi- cally ct p � 24�m. In experiments Ref. 4 a unique phenomenon was ob- served: when one short current pulse was given to the heater, two-phonon pulses, well separated in time, were detected. The first comprised l-phonons, and the second, h-phonons. From subsequent experiments (see Refs. 5, 6) it was unambiguously demonstrated that the h-phonon pulse was not injected by the heater but was created by the l-phonon pulse during its motion from the heater to the de- tector. The theory of this surprising phenomenon, when a cold l-phonon pulse with temperature close to 1 K, creates high-energy phonons with energy � � 10 K, was given in Refs. 7–9. In Refs. 7–9 it was shown that l-phonon pulse is in equi- librium due to the fast 3pp. The h-phonons are created within the l-phonon pulse by slow 4pp with rate �4 pp . The h-phonons have a group velocity u h � 189 m/s that is smaller than the velocity of the l-phonon pulse c � 238 m/s. The difference in these velocities and the relatively weak in- teraction between h- and l-phonons (see Eq. (2)) leads to the h-phonons leaving the l-phonon pulse through its rear wall and forming a pulse of noninteracting h-phonons which ar- rives at the detector after the l-phonon pulse. However the theory created in Refs. 7–9 gives no ex- planation for some of the phenomena observed in experi- ments. One of these is that the measured solid angle in mo- mentum space occupied by the h-phonons is much smaller than the solid angle occupied by l-phonons. Also it was difficult to explain how the l-phonon signal amplitude de- pended on heater power because the occupied solid angle in momentum space was constant during the propagation. The theory predicted that all pulses cool to about the same temperature due to the creation of h-phonons; the final temperature is where the h-phonon creation rate is negli- gible. So if the temperature and occupied solid angle are the same then the signal on the detector from the pulses of different powers should be the same. But experiments showed that the l-phonon signal amplitude increased with increasing of heater power, so it was proposed that larger solid angles were created at higher powers [10]. In this paper the exact local-equilibrium distribution function, for phonons in a pulse, is used. It allows us to obtain the angular dependence of the h-phonon creation processes, and to explain the above mentioned experi- mental phenomena. In Refs. 7–9 a more simple approxi- mation of the distribution function was used, the cone ap- proximation, in which a cone of occupied states is cut from an isotopic distribution at tempeature T p . Some of results presented here have been published [11]; here we expand and extend those results, and from this foundation obtain several new results. In particular, the important case of creation and decay of high-energy phonons with momentum directed along the anisotropy axis of phonon system, is investigated for the first time. For this case, an exact expression for the creation and de- cay rates in an anisotropic system is obtained, together with the relaxation rate of the isotropic phonon system. Hence we derive an explicit analytical expression for the rate of h-phonon creation along the anisotropy axis. We also construct an analytical approximation for the rate in range of angles where the collision integrals are not negli- gible. This allows us to analyse the rate dependences on all the parameters of the problem. It is shown that the h-phonon creation rate strongly depends on the momen- tum pd , which separates the l- and h-phonon subsystems. From a comparison of the theory, developed in this paper, with experimental data, we conclude that p pd c� . 2. The kinetic equation for the distribution function of h-phonons In this section we consider the four-phonon processes (4pp) which determine the creation rate of h-phonons in anisotropic and isotropic l-phonon systems. The conser- vation laws of energy and momentum for 4pp can be written as � � � �1 2 3 4 1 2 3 4� � � � � �, .p p p p (3) Hereinafter, the phonon with energy �1 and momentum p1 is the h-phonon and the other three phonons are l-phonons. This is the type one 4pp scattering [12] and is the most impor- tant process for h-phonon creation. The kinetic equation describing the rate of change of the distribution function, n n( )p1 1� , due to 4pp can be written as dn dt N Nb d 1 1 1� �( ) ( ),p p (4) 524 Fizika Nizkikh Temperatur, 2007, v. 33, No. 5 I.N. Adamenko, Yu.A. Kitsenko, K.E. Nemchenko, V.A. Slipko, and A.F.G. Wyatt where N b ( )p1 and N d ( )p1 are, respectively, the rates of increasing and decreasing numbers of h-phonons in the state with momentum p1, due to collisions. N b d, ( )p1 can be written as N b d, ( )p1 � � �W n d p d p d pb d( , ; , ( ( .,p p p p p1 2 3 4 3 2 3 3 3 4�� �� � �� � (5) Here W ( , ; , )p p p p1 2 3 4 defines the probability density of the process (3), the �-functions correspond to the conser- vation of energy and momentum; � � � � �� � � � �1 2 3 4 and p p p p p� � � � �1 2 3 4 . The factors nb d, are n n n n nb � � � 3 0 4 0 2 0 11 1 ( ) ( ) ( ) ( )( ), n n n n nd � � �1 2 0 3 0 4 0 1 1 ( ) ( ) ( ) ( )( ), (6) where n n l l ( ) ( )( ) 0 0� p is the equilibrium distribution function of l-phonons which are considered to be instan- taneously in equilibrium due to three-phonon processes. The equilibrium distribution function of l-phonons as intoduced in Refs. 1, 2 can be written as n k Tl l l B ( )( ) exp0 1 1p p u � � � � � � � � � � � � � � �� . (7) The drift velocity u can be written as u N� �c ( ),1 ! (8) where N is the unit vector directed along the direction of the total momentum of the l-phonon system; this defines the anisotropy axis of the phonon system. This equation defines !, the anisotropy parameter. In weakly aniso- tropic systems, ! is close to unity. For strongly ani- sotropic systems, which corresponds to the experiments [4–6,13,14], ! �� 1. In all further calculations it will be more convenient to us to use the expression n p cp k T l l l B l l l ( )( , ) exp ( )0 1" � ! " " !� � � � � � �� � � �� � � � � � � �1 , (9) where " l l lp� �1 p N / . Equation (9) was obtained by substitution of Eqs. (1) and (8) into (7). In an earlier description of anisotropic phonon systems we used the approximate distribution function (see Refs. 7–9) which can be written as n k Tp l p l l B p ( ) /( ) ){exp( ) }0 11p � � � �#�$ $ � . (10) This distribution function has a simple physical mean- ing and includes all the parameters of anisotropic phonon systems: temperature T p and the anisotropy which is de- fined by the cone angle $ p . The values of T p and $ p can be estimated from the experimental data. However the de- scription of the creation of h-phonons, at any angle to the anisotropy axis (see, for example, [1,2]), can only be made using the exact local-equilibrium distribution func- tion (7) for the l-phonons. As that is the purpose of this pa- per, we shall use only the parameters! and T from now on. The relationships between the parameters !, T and $ p , T p come from the equating energy and momentum in the two descriptions, and were obtained in Refs. 1, 2. The interaction of phonons in superfluid helium is de- scribed by the Landau Hamiltonian (see, for example, [15]), which we write as � � � � .H H V Vph � � �0 3 4 (11) Here �H 0 is the Hamiltonian of noninteracting phonons. The terms �V3 and �V4 describe the interaction of phonons to the third and fourth orders of small deviations of a sys- tem from an equilibrium state, respectively. The probability density of four-phonon process fol- lowing Refs. 9, 15, 16, can be written as W V Hfi( , ; , ) | | ( ) p p p p1 2 3 4 2 2 6 2 1 2 � % %� � . (12) Here V is the volume of system and Hfi is the amplitude of four-phonon process which is obtained by second or- der perturbation theory on �V3 and first order perturbation theory on �V4 with a help of standard procedures (see, for example, [9,15–18]). Hence H V V E E Vfi iQ � & '& ' � � &( p p Q Q |p p p p p Q 3 4 3 3 2 3 4 4 , | � | | � , | � |1, 1, p2', (13) where Q is an intermediate state with energy EQ, and Ei is the energy of initial state. The interaction of phonons with momenta p1 and p2 to create phonons with momenta p3 and p4 , has six interme- diate states I–VI in which phonons have momenta I. II. III.p p p p p p p p p p1 2 2 3 1 3 2 4 1 4� � �; , , ; , , ; IV V.. , , ; , , ;p p p p p p p p1 3 2 3 1 4 2 4� � VI. p p p p p p1 2 3 4 1 2, , , , .� � (14) From expressions (13) and (14) we obtain H p p p p V Mfi � 1 2 3 4 8) � , (15) where ) �145 kg/m 3 is the density of He II, M M M M M M M M� � � � � � � �( ) ( ) ( ) ( ) ( ) ( )1 13 2 14 3 23 3 24 3 5 4 (16) is the matrix element for 4pp, which consists of seven terms, six of which correspond to the six intermediate states (14), and the seventh, to first order of perturbation theory on �V4 . Creation of high-energy phonons by four-phonon processes in anisotropic phonon systems of He II7 Fizika Nizkikh Temperatur, 2007, v. 33, No. 5 525 The superscripts show the number of phonons in the inter- mediate state. We can write the matrix elements as M u( ) ( )1 1 2 1 2 1 2 1 2 1 1 2 2 1 22 1� � � � � � � � � � � � � � � n n n n n n � � � �� �( )2 1 3 4 3 3 4 4 3 4u n n n n n n , (17) M u( ) ( )5 1 2 1 2 1 2 1 2 1 1 2 2 1 22 1� � � � � � � � � � � � � � � � n n n n n n � � � �� �( )2 1 3 4 3 3 4 4 3 4u n n n n n n , (18) M u w4 24 1� � �{( ) }, (19) M u 13 3 1 3 1 3 1 3 1 3 1 1 3 3 1 32 1 ( ) ( )� � � � � � � � � � � � � � � n n n n n n � � � �� �( )2 1 2 4 2 1 3 4 1 3u n n n n n n . (20) Other terms of expression (16), i.e., M 23 3( ) , M 14 3( ) , M 24 3( ) can be obtained from M 13 3( ) by permuting the corres- ponding subscripts. Here n pi i ip� / , � � � �i i� p , u c� ( / )) ( / ) .* * �c ) 2 84 is the Gr��uneisen constant and w � ( / )( / ) .) )2 2 2 0 188c c* * � . The matrix elements are given in detail in the Appendix A. We note, that first three terms of Eq. (16) are resonant: when � �( p � 0 their denominators can vanish, which leads to an essential divergence of the matrix element. These three terms give the main contribution to Eq. (16). The rest of the terms give small contributions of the same order of magnitude. We define the rate of the 4pp scattering by �1 1 3 2 3 3 3 4( )p � � d p d p d p � �W n n n( , ; , ) ( ( ( )( ) . ( ) ( ) ( ) p p p p p1 2 3 4 2 0 3 0 4 0 1 1� �� � �� � (21) Then from Eq. (5), the expression for N d can be written as N nd � 1 1� . (22) For N b we use the equality n n n k T n n B 3 0 4 0 2 0 1 1 2 0 1 1 ( ) ( ) ( ) ( ) ( ) (� � � �� � �� � � �� �exp � p u 3 0 4 0 1 ( ) ( ) )( ),� n (23) which follows from the Eqs. (3) and (7). Then we can write N k T nb B � � �� � �� � � �� �exp � ��1 1 1 11 p u ( . (24) For short pulses that were used in experiments [13,14], n1 is much less than the unity. (We shall only con- sider short pulses from now on.) Therefore the relation N k Tb B ( )sh exp� � �� � �� � � �� � �1 1 1 p u (25) is always satisfied in short pulses. To see the important role of "1 in determining the cre- ation rate of h-phonons, we rewrite the exponent in (25): � � � � � � � ( ) ( ) � � ! " " !1 1 1 1 1 1 p u k T cp k TB B . (26) For h-phonons � 1 is small and negative: when 10 111K � �~p K then 0 0 023 �� . . For the anisotropic phonon systems studied experimentally, the energy densi- ties in the liquid helium and the angular distribution of the l-phonons are consistent with ! � 0.02 and temperature T � 0.04 K. These values of T � 0 041. K and ! � 0 02. in the distribution function (9) correspond to the same energy density and momentum density as the values T p �1 K and $ p � +12 in the cone distribution function (10). Although ! and � are small, "1 is not necessarily small; it can vary from 0, when p1 is along the anisotropy axis, to 2, when p1 is antiparallel to the anisotropy axis. When"1 is nearly zero then the modulus of the exponent is not large, and the expo- nential term in (25) is not so small, and the creation rate is not small. However when " !1 and | |� , then the expo- nent is large and negative, and the exponential term in (25) is very small, which makes the creation rate negligible. In order to calculate the creation rate N b and the decay rate N d , which appear in the kinetic equation (4), we must find the rate �1 which is done in the next section. 3. The rate of h-phonons creation and decay in the anisotropic l-phonon system 3.1. The general expression for the rate �1 To derive the general expression for the rate �1, we re- write the expression (21), taking into account (12)–(15), in spherical coordinates � % ) , " , " , "1 1 11 5 7 2 2 2 2 3 3 3 4 4 4 2 3 3 3 4 2 � � p dp d d dp d d dp d d p p p � 3 � �M n n n� � � 2 2 0 3 0 4 0 1 1� �� � �( ( ( )( ) ( ) ( ) ( ) p , (27) where " i i ip� �1 p N / . Without any restriction on generality, we can choose the angle ,1 as the computing origin of angles , i . In this case the �-functions can be written as �� � � , , , �p� � � � � - - - -( cos cos cosp p p p1 2 2 3 3 4 4 � � - - -� , , , �( sin sin sinp p p2 2 3 3 4 4 � � �� ( )|| || || ||p p p p1 2 3 4 . (28) 526 Fizika Nizkikh Temperatur, 2007, v. 33, No. 5 I.N. Adamenko, Yu.A. Kitsenko, K.E. Nemchenko, V.A. Slipko, and A.F.G. Wyatt � � � � .( ( )� � � � � � 1 1 2 3 4 c p p p p , (29) where p pi i i i- � �2 2" " , p pi i i|| ( )� �1 " , . � � � �� � � �p p p p3 3 4 4 1 1 2 2. (30) We note that the value ., when p1 is an h-phonon, is al- ways positive (. 0). In order to make the integration with respect to , 3 and , 4 we introduce new variables X p p� �- -3 3 4 4cos cos ,, , Y p p� �- -3 3 4 4sin sin, , . (31) Then the expression (27) can be written as � % ) " " " , 1 1 10 5 7 2 2 3 3 3 4 3 2 3 4 2 3 4 2 2 4 � p p p p dp dp dp d d d d dXdY p� 3 2 4 2 2 2 3 2 4 2 2 - - - -� � � � � p X Y p p( ) � �M n n n� � � 2 2 0 3 0 4 0 1 1� �� � �( ( ( )( ). ( ) ( ) ( ) p (32) Hereinafter the integration is made so that the radicands are not negative. As M � depends on cos, 3 and cos, 4 , to make the inte- gration it is necessary to solve the system of the equations p p p p p p 1 2 2 3 3 4 4 2 2 3 3 0- - - - - - � � � � � cos cos cos , sin sin , , , , , � � � � � -p4 4 0sin ,, (33) with respect to cos, 3 and cos, 4 . The system of equations (33) has two solutions which can be written as cos ( cos ) ( ) sin( ), , , 3 1 2 2 3 2 4 2 2 2 32 / - - - - -� � � � /p p A p p p R Ap - , (34) cos ( cos )( ) sin( ), , , 4 1 2 2 3 2 4 2 2 2 42 / - - - - -� � � �p p A p p p R Ap � - , (35) where A p p p p� � �- - - -1 2 2 2 1 2 22 cos ,, (36) R p p p p p p p p� � � � � �- - - - - - - -4 23 2 4 2 1 2 2 2 3 2 4 2 1 2 2 2( cos ) ., (37) At first we integrate with respect to X and Y , and then with respect to p4 and " 4 with the help of the �-functions. As a result we have � " " ,1 1 2 3 2 3 2 2 3 3 3 4 2 � �Kp dp dp d d d p p p R � � �� �{ } ( )( ),( ) ( ) ( ) ( ) ( ) M M n n n2 2 2 0 3 0 4 0 1 1 (38) where K c � 1 210 5 7 2% )� , (39) M M( ) ( ) ( ) (cos cos , cos cos ),/ / /� � �� , , , ,3 3 4 4 (40) p p p p4 1 2 3� � � �., (41) " " " " . 4 1 1 2 2 3 3 4 � � � �p p p p . (42) Further integration cannot be precisely made analyti- cally because of the complicated integrand expression. Therefore we first consider the important case of "1 0� and secondly present some dependences of the rate �1 ob- tained from Eq. (38). 3.2. The rate �1 for h-phonons moving along the anisotropy axis This case is important as it is possible to obtain an ex- plicit analytical expression for �1. In Ref. 9 the rates of four-phonon processes were calculated when "1 0� using the approximate cone distribution function (10). When "1 0� , the expression (38) becomes much sim- pler as firstly the dependence of matrix element on, 3 and , 4 vanishes, and secondly the dependence of the integrand expression on , 2 disappears. So the integration over , 2 can be easily made analytically. Hence we find � % " " " " " "" " 1 1 2 3 2 3 3 0 3 3 3 0 2 1 1 � � � � � � ��Kp dp dp d d ( ) ( ) � � � � � � � � � � � 1 2 1 2 2 1 2 2 1 2p p p p " ( ) � � ��p p p p p M n n n2 3 3 2 4 2 1 2 0 2 2 0 3 0 4 01 1 1( ) ( )( ) , ( ) ( ) ( ) � " (43) where M M M� " " "1 0 1 10 0 � � �� � � �( ) ( )( ) ( ) , (44) " " 3 1 0 / � � � / � � �� � b p p p a " " " " " " " "1 1 0 1 2 2 3 2 2 2 2 2 2 0 8 2 2 ( )( )( ) | min max | . (45) Here Creation of high-energy phonons by four-phonon processes in anisotropic phonon systems of He II7 Fizika Nizkikh Temperatur, 2007, v. 33, No. 5 527 b p p p p p p p p p p " " .1 0 1 2 2 3 2 2 2 3 1 4 2 3 18 4 2 2 2 � � � � � � �( � � � � �2 4 24 2 2 3 1 2 4p p p p p. . �" . � .( ( ) , (46) a p p p p p p p " " 1 0 3 2 1 2 2 1 2 2 1 2 2 4 1 2� � � � � � � � � � � � � � ( ) ( ) , (47) " . . 2 3 4 1 2 2 2 2 min ( ) ,� � �p p p p (48) " . . 2 3 4 1 2 2 2 2 max ( )( ) .� � �p p p p (49) We note that Eq. (43) for �1 is very important as it not only describes the rate of h-phonon creation at $1 0� but also gives the relaxation in isotropic phonon system, i.e., when ! �1, because in the isotropic case there is no de- pendence on $1 and every direction is identical to $1 0� . Following Ref. 9, we can replace M � "1 0� with M which can be written as M u m m m u m� � � � � �4 1 4 12 2 3 4 2( ) ( ) ( ) . (50) Here m p p p p 2 1 2 2 1 2 2 2 1 � � � ( ) , " 0 (51) m p p p p 3 1 3 2 1 3 3 3 1 � � � ( ) , " 0 (52) m p p p p 4 1 4 2 1 4 4 4 1 � � � ( ) , " 0 (53) where 0 i are functions dependent on the dispersion: 02 1 2 1 2 1 2 1 2� � � � � p p p p f f f( ) , 03 1 3 1 3 3 1 3 1� � � �� p p p p f f f( ), (54) 0 � � �4 1 4 1 4 4 1 4 1� � � � �� p p p p f f f f p pi i i( .) and (55) For all further calculations it is necessary to have an analytical approximation for the function � �( p . In this section and further on, we shall use the approximation of function � �( p , which was obtained in [9] � � 1 2 ( , , p c u c p p p p p p p p p c u c c c c f c c � � � � � � � � � � � � � � � 2 1 2 2 2 2 � 3 � 4 56 7 89 � � � � : : � : : u c p p p p pc c f 2 ( ) , .� (56) Here ~pc �10 K, ~ .p f � 8 26 K, ~p� � 27 K, 1 �176. , 2 �113. , u pc p pc � * * ��� / | /189 m s, k u cB c3 �/ � ( / )k cB ( / )|* �� 2 2� *p p pc – 19.8 m/(s;K). Further integration can be made by replacing, in the slowly varying functions of momenta and angles in the integrand of Eq. (43), the momenta and angles by their typical values. As a result we have � � 1 0 2 1 1 2 3 2 1 2 � � � �� � � �� � � � K T cp k T m p p p p p B c cexp ( , ) � � � �� � � �� � � � �p p p p p p p pc c c c1 1 2 1 1 52 12 6 / ( ) ( ( ))< < < . (57) Here K u c u k c c B 0 4 4 3 7 2 2 1 2 � � � � � � � � � ( ) ( ) , % )� (58) m m p p p p p p p p � � � � � � � �� � � ��" . " . 2 1 2 1 2 3 4 2 1 1 2 , ( ) ,, (59) < � ! � �� � � � c k T t B t ( ) . .2 1 2 0 02, where (60) The result (57) allows us to analyze the dependences of the rate �1 on p1, T , ! and the parameters of liquid helium. So for example, the fast decrease of �1 with increasing p1 is mainly caused by the exponent. We see from (57) that the rate is proportional to the minus one third power of the dispersion as� 0. Moreover the form of result (57) will be used below when obtaining an approximation for the rate �1 in the range of not too-large "1, which is mainly re- alized in a pulse. We note that Eq. (57) is in good agree- ment with the results of exact numerical calculation of the rates from Eq. (43) near T � 0 041. K and ! � 0 02. . At other values of ! and T the numerical agreement with the rates calculated from Eq. (43) are not so good, but the dependences are qualitatively the same. We note, that in obtaining the result (57), we used the fact that the main contribution, in integration over p3, is due to phonons with momenta close to pc , which we have assumed separates the l- and h-phonon subsystems. This leads to the conclusion that the rate of h-phonon creation strongly depends on the numerical value of momentum which separates the l- and h-phonon subsystems. The pos- sibility that this momentum, pd , which separates the l- and h-phonon subsystems, is not equal to pc will be discussed in section 6. 3.3. Dependence of the rate �1 on all parameters The dependences of �1 on p1, "1, T and !, can only be obtained by numerical integration of Eq. (38). The results of this are shown in Fig. 1: in Fig. 1,a is shown the angular 528 Fizika Nizkikh Temperatur, 2007, v. 33, No. 5 I.N. Adamenko, Yu.A. Kitsenko, K.E. Nemchenko, V.A. Slipko, and A.F.G. Wyatt dependence of the rate �1 at different values of p1 and fixed values of T � 0 041. K and ! � 0 02. , which correspond to the conditions of the experiments [13,14]. We see that there is a maximum when ~p1 13� K. The appearance of this maximum at ~p1 10� K is caused by several reasons: firstly it follows from the conservation laws of energy and momentum that the four-phonon processes are prohibited if the angle between the momenta of phonons p1 and p2 is less than " $ .12 12 1 2 1 2 1min mincos� � = �p p p p . (61) Secondly the distribution function of phonons, in a pulse, has a sharp maximum at $2 0� . Thus, the majority of phonons in a pulse have momentum p2 directed along the anisotropy axis of the system. However at $1 0� phonons with momentum p1 cannot interact with phonons having momentum p2 directed along the anisotropy axis because there is a minimum angle, $12min , for interaction between them. Thus, when $1 0� , the interaction of phonons with momentum p1 with most of the phonons of the pulse is prohibited by the conservation of energy and momentum. With increasing angle $1, an increasing number of phon- ons in the pulse interact with the h-phonon. This leads to the increasing rate in the initial part of the curve, up to $1 40� +, see Fig. 1,a. The following decrease in the rate is caused by the decreasing value of the squared matrix element with increasing angle $1. At these larger angles, the h-phonon with momentum p1 can interact with almost all phonons of the pulse. The squared matrix element is approximately constant for $ $1 1 80� = + ME and even could be slightly increasing with the growth of $1 in this range of angles. This is due to the cancellation of terms in (16). Then the slow growth of the rate at these angles is determined by the factor R in the integrand of Eq. (38) and by the angular dependence of matrix element, see Fig 1,a. We note that it is important to retain all terms in M � , if only one term is considered there are serious qual- itative errors; for example for M ( )1 , M 13 3( ) or M 14 3( ) , the rate �1 decreases with growth $1, instead of increasing. The situation is different when the p1 is far from pc . From (61) we see that angle $12min increases with larger p1 and becomes greater than the angle $1ME . Then the rate rises monotonically with increasing $1 and there is no range of angles where the rate decreases. This behavior is apparent when ~p1 13� K, see Fig. 1,a. In Fig. 1,b is shown the momentum dependence of the rate �1 at different values of $1 (0 � , 30 � , 60 � , 90 � and 180 + corresponding to curves 1–5, respectively) with the same values of T and ! as in Fig. 1,a. It can be seen, that with increasing angle, the rate �1 is higher in more of the momentum range, and at$1 180� + the dependence on mo- mentum is actually absent. In Fig. 1,b we also see that the curves cross each other. The physical reasons for these be- haviors are the same as those discussed above. We note that momentum dependence of �1, when $1 0� , was ob- Creation of high-energy phonons by four-phonon processes in anisotropic phonon systems of He II7 Fizika Nizkikh Temperatur, 2007, v. 33, No. 5 529 10 8 a1 2 3 4 5 6 10 7 10 6 10 5 10 4 10 3 10 2 10 1 0 20 40 60 80 100 120 140 160 180 b 10 11 12 13 14 1' 1 2 3 4 5 10 8 $ , degree 1 cp /k , K1 B 1' 2' 3' 10 3 10 2 10 1 10 7 10 6 10 5 10 4 � 1 , s– 1 � 1 , s– 1 Fig. 1. a — The rate �1 is shown as a function of the angle $1 between p1 and the propagation direction calculated from Eq. (38) for different values of ~p1 equal to 10 (1), 10.5 (2), 11 (3), 12 (4), 13 (5), and 14 (6) K. The dashed lines are corresponding approximate rates calculated from Eq. (62). b — The dependence of the rate �1 on momentum cp kB1 / cal- culated from Eq. (38) for different $1 equal to 0 � (1), 30 � (2), 60 � (3), 90 � (4), and 180 + (5). The dashed lines are the corre- sponding approximate rates calculated from Eq. (62). All calculations had T � 0041. K and ! � 002. . 0.020 0.025 0.030 0.035 0.040 T, K 1 2 3 4 5 10 7 10 6 10 5 10 4 10 3 � 1 , s– 1 Fig. 2. The temperature dependence of the rate � "1 1 1 0( , )p pc� � for different values of the anisotropy parame- ter ! equal to 0.01 (1), 0.015 (2), 0.02 (3), 0.025 (4), and 0.03 (5). The dashed lines represent the same dependences calculated from Eq. (62). tained in Ref. 9 with the approximate distribution func- tion (10); the rate is similar to that presented here. In Fig. 2 the dependence of �1 on T for different values of ! is shown for ~p1 10� K and $1 0� . From Fig. 2 it can be seen, that the rate increases with both increasing temperature and with decreasing values of the anisotropy parameter !. Such behavior is completely determined by the dependence of distribution function n 2 0( ) on the parameters ! and T . To solve a number of problems it is convenient to use analytical expression for �1. However it is impossible to obtain an analytical expression for � " !1 1 1( , , , )p T from the fivefold integral (38). In this case it is useful to have an approximate expression for � " !1 1 1( , , , )p T . Using the data from the numerical calculation of the dependences of the rate on p1 and "1, which was made with a help of Eq. (38) at different values of ! and T , then least squares fitting and taking into account (57), it is pos- sible to obtain a numerical approximation for the rate �1 on all parameters of the problem, which can be written as � " ! !1 1 1 1 1 221 39 10 1 1 2 1( , , , ) . ( ) ( ) p T T p p� ; ; � � � � � � � � � � e ) 10 24 22 4 1 1 1 1 3 1 5 11 � 1 ! 11 "T p p p T pc p . ( ) ( ) ( ( ) , (62) where T should be substituted in Kelvins and 1 i should be written as 1 11 1 1 2 1 110 37 6 06 3 11 2 53( ) . . ( ) . .p p p p p pc c � � � �, , (63) 1 3 1 116 778 18 456( ) . .p p pc � � , (64) 1 4 1 1 0 668 24 29 95( ) exp . . p p p p c c � �� � �� � � �� � � � � � � � � , (65) 1 5 1 16 62 5 71( ) . .p p pc � � . (66) The approximation (62) can be used for 0 01 0 06. .� �! and 0 016 0 06. .K K� �T . We note that all experiments were carried out in this range of values of ! and T . Equa- tion (62) is valid in the momentum range of ~p1 from ~pc up to 11 K, where the rate N b ( )sh is significantly above zero. We notice, that approximation (62) is applicable in the range of angles where N b ( )sh is large. This is due to the fact that �1 in N b ( )sh is multiplied by exp[ ( ]� ��1 p u) /1 k TB , which very quickly decreases with increasing angle, and makes N b ( )sh relatively small at angles $1 10� +, see Fig. 3,a. The comparison of the numerical calculations of the rate with the approximation (62) one can see in Figs. 1 and 2 where the dashed curves were obtained with a help of Eq. (62). We see that for the given range of variables, the approximation (62) is rather good. 4. Angular distribution of l- and h-phonons The angular distribution of l- and h-phonons can be de- scribed by the probability density of the angular distribu- tion of phonons. The probability density of created h-phonons as a function of "1 is defined by expression W T N p T p dp E h p p b c( , , ) ( , , , ) � max ( ) " ! � " ! % 1 1 1 1 1 2 1 2 34 � � sh � h , (67) where � ( , , , ) max ( ) E d N p T p dph p p b c � � � 1 4 2 3 1 0 2 1 1 1 1 2 1 % " � " ! � sh (68) is the total energy of h-phonons, created in unit time and unit volume, N p T b ( ) ( , , , ) sh 1 1" ! is defined by Eq. (25), and the upper limit of integration over momentum is cho- sen to be p pcmax .�14 . The value of the integrals are not sensitive to the exact value of this upper limit as long as it is well above pc : even if p pcmax .�11 then the results 530 Fizika Nizkikh Temperatur, 2007, v. 33, No. 5 I.N. Adamenko, Yu.A. Kitsenko, K.E. Nemchenko, V.A. Slipko, and A.F.G. Wyatt a1 2 3 4 5 0 2 4 6 8 10 12 14 16 18 20 b1 2 3 10 11 12 13 14 $ , degree 1 cp /k , K1 B b N , s (s h ) – 1 10 5 10 4 10 3 10 2 10 1 10 0 10 5 10 4 10 3 10 2 10 1 10 0 N , s (s h ) – 1 Fig. 3. a — The creation rate of h-phonons, N b ( )sh in short pulses, is shown as a function of the angle $1 between p1 and the propa- gation direction for different values of cp kB1 / equal to 10 (1), 10.5 (2), 11 (3), 12 (4) and 13 (5) K. b — The dependence of N b ( )sh on momentum cp kB1 / for different values of $1 equal to 0�(1), 6+(2), and 12+ (3). All curves have ! � 002. and T � 0041. K. change very little. The main contribution to the integrals is due to h-phonons with momentum close to pc because the function N p b ( ) ( ) sh 1 rapidly decreases when p1 in- creases, see Fig. 3,b. Similarly we can also introduce the probability density of l-phonons as a function of " l W T n p T p dp E l l l p l l l l l l c ( , , ) ( , , , ) ( ) " ! � " ! % � � 0 0 2 2 34 � , (69) where E d n p T p dpl l l p l l l l l c � � � 1 4 2 3 0 2 0 0 2 % " � " ! � ( ) ( , , , ) (70) is the total energy of l-phonons in unit volume of the pulse, and n p T l l l ( ) ( , , , ) 0 " ! is defined by Eq. (9). The results of numerical calculations of the depend- ence ofWl andWh on $ "l h l h, ,( )� �arccos 1 , obtained from Eqs. (67) and (69) at ! � 0 02. and T � 0 041. K, which are values for t � 0, are given in Fig. 4. We see that Wh is a considerably sharper function angle thanWl . For ! � 0 061. and T � 0 058. K, which are values after 42 �s, we see that W tl ( )� 42�s is even wider than W tl ( )� 0 . The sharpness of functionsWl h, is defined by the angu- lar width of the corresponding distributions which are given by relation " " "l h l h l h l hW d, , , ,� � 0 2 . (71) Figure 5 shows the dependences of $l h, � arccos ( ),1�" l h on T , at different values of !. It can be seen that the in- equality $ $h l� is always satisfied. Thus the created h-phonons always occupy a narrower cone than the l-phonons. This characteristic was observed in experi- ments [5] and was called «the concentration of h-phonons near the anisotropy axis of the system». Experiments showed, that h-phonons are in a cone with a cone angle about 4+ , while l-phonons occupy a cone with a cone angle close to 12 +. We notice, that results of calculations with values of parameters which are typical for experiments [13,14], T � 0 041. K and ! � 0 02. , give $l = +12 and $h = +5 5. near the heater, and near the bolometer we have $ �l t( ) .� = +42 16 8s and $h is not expected to change much. These calculated values after 42 �s are similar to experimental values. Thus, the reasons for the observed concentration of h-phonons near the anisotropy axis of the system, can now be explained as follows. The rapidly decreasing value ofWh with increasing "1, is related to the angular dependence of N b ( )sh which is shown on Fig. 3. It can be seen in Fig. 3,a, that N b ( )sh very quickly decreases with increasing $1. This decrease is de- termined by the multiplier exp )[ ( / ]� ��1 1p u k TB in ex- pression (25), which has a sharp maximum at $1 0� , see Eq. (26). While in Fig. 3,a the presence of the exponential multiplier leads to the fast decrease of the function N b ( ) ( ) sh $1 while function � $1 1( ) is increasing. In Fig. 3,b the presence of the exponential multiplier makes the mo- mentum dependence of N p b ( ) ( ) sh 1 more flat than the de- Creation of high-energy phonons by four-phonon processes in anisotropic phonon systems of He II7 Fizika Nizkikh Temperatur, 2007, v. 33, No. 5 531 W 120 100 80 60 40 20 0 0 3 6 9 12 15 18 21 24 27 30 $, degree W (t = 0 )h �s W (t = 42 s)�1 1W (t = 0 s)� Fig. 4. The angular dependence of the probability density W t( )� 0 for the h- and l-phonon distributions at ! � 002. and T � 0041. K obtained from Eqs. (67) and (69), respectively. Also W tl( � 42�s) is shown for ! � 0061. and T � 0058. K which are the corresponding values at t � 42�s. Note the angular distribution is narrower for h-phonons than for l-phonons. a 3 2 1 T, K 0.015 0.020 0.025 0.030 0.035 0.040 $ , d eg re e 1 – 13 12 11 10 9 b 1 2 3 T, K 0.015 0.020 0.025 0.030 0.035 0.040 – 6.0 5.8 5.6 5.4 5.2 5.0 4.8 4.6 4.4 $ , d eg re e h Fig. 5. The angular widths of the l- and h-phonon distributions, $l (a) and $h (b), are shown as functions of temperature T, for different values of ! equal to 0.01 (1), 0.02 (2), and 0.03 (3). pendence �1 1( )p . This can be explained by the fact that the exponential exp )[ ( ]� ��1 1p u / k TB quickly increases with increasing momentum p1, at fixed values of $1, due to the increase in the absolute value of the negative func- tion � �( p1 with increasing p1, see Eq. (26). An important characteristic of h-phonon creation is the relation E El h/ � . It is the characteristic time it takes for the energy of the created h-phonons to become equal to the initial energy of l-phonon pulse, if the parameters ! and T of l-phonon pulse are assumed constant. This characteris- tic time shows the intensity of the h-phonon creation process. Comparing this time with the time of the pulse propa- gation > �p � = ; ��10 238 4 2 10 425mm m s s s / . , (72) we can determine whether l-phonons have time to trans- fer a significant part of their energy to h-phonons during their propagation from the heater to the detector. Figure 6 shows the dependence of E El h/ � on temperature T for different values of !. One can see that when ! � 0 01. , the h-phonon creation is rapid at temperatures T 0 02. K, when! � 0 02. atT 0 032. K, and when! � 0 03. there is prac- tically no h-phonon creation at any temperature. 5. Evolution of a short l-phonon pulse caused by h-phonon creation A short phonon pulse is such that all the created h-phonons are lost from the l-phonon pulse well before equilibrium between l- and h-phonons is established. The kinetic equation (4) can be written as dn dt N b 1 � ( )sh . (73) We multiply the right and left part of Eq. (73) by �1 and integrate over all d p3 1 32/ ( )%� . As a result we have d dt Eh h � � � , (74) where �Eh is defined by expression (68), and � % " �h p p d dp p n c � � � 1 4 2 3 1 0 2 1 1 2 1 1 � max . (75) From conservation of energy it follows that d dt dE dt h l� � � , (76) where El is defined by expression (70). From (74) and (76) we have � � dE dt El h � . (77) Now we multiply the right and left part of Eq. (73) by p z1 and in a similar way obtain � � dP dt Pl h � , (78) where P d n p T p dpl l l p l l l l l c � �� � 1 4 1 2 3 0 2 0 0 3 % " " ! " � ( ) ( , , , ) ( ) , (79) � ( , , , ) ( ) ( ) max P d N p T p dh b p p c � �� � 1 4 1 2 3 1 0 2 1 1 1 3 1 % " " ! " � sh p1. (80) Thus, we have obtained two equations (77) and (78) that must be solved with respect to functions ! �( t and T t( ) with the initial conditions ! � !( t � �0 0 and T t T( )� �0 0. The solution of combined equations (77) and (78) is dis- cussed in detail in the Appendix B. Figure 7,a represents the temperature T t( ) dependence on time calculated with T0 0 041� . and ! 0 0 02� . , 0.025, 0.03, 0.035 (curves 1–4, respectively). In Fig. 7,b is shown the anisotropy parameter ! �( t as a function of time, for ! 0 0 02� . and T0 0 025� . , 0.030, 0.036, 0.041 (curves 1–4, respectively). It can be seen from Fig. 7 that h-phonon creation leads not only to increasing ! but also to increasing T ; nevertheless the total energy density of l-phonon pulse decreases with time (see Fig. 8). In Fig. 8 the time dependence of the l-phonons energy density E t E t T tl l( ) ( ( , ( ))� ! � and the l-phonon pulse an- gular width $ $ ! �l lt t T t( ) ( ( , ( ))� are shown. The broaden- ing of a pulse with time, seen in Fig. 8, is due to the con- 532 Fizika Nizkikh Temperatur, 2007, v. 33, No. 5 I.N. Adamenko, Yu.A. Kitsenko, K.E. Nemchenko, V.A. Slipko, and A.F.G. Wyatt T, K 3 2 1 10 10 10 10 10 10 0.015 0.020 0.025 0.030 0.035 0.040 >p 3 1 –1 –3 –5 –7 E /E , s l h. Fig. 6. The characteristic time for l-phonon energy to be con- verted to h-phonon energy is given by the ratio E El h/ � is shown as a function of temperature T for different values of ! equal to 0.01 (1), 0.02 (2), and 0.03 (3). The dashed line shows the time for a pulse to travel 10 mm, i.e. 42 �s, from Eq. (72). We see more energy is converted when ! is small, i.e., most anisotropic. servation of momentum when h-phonons are created in the l-phonon pulse. The created h-phonons, as can be seen in Fig. 4, have a sharper angular dependence and that is why they carry away momentum mainly parallel to the anisotropy axis. To compensate for this, the l-phonon an- gular distribution must broaden. For other values of ! 0 and T0, all the dependencies are qualitatively the same as shown in the figures. We introduce ? �( ( , ) ( ( , ( )) ( , t E T E t T t E T l l l � �! ! � ! 0 0 0 0 ) , (81) which shows the fraction of the l-phonon energy which transforms into h-phonons. In Fig. 9 the function ? �( t is shown at ! 0 0 02� . for five values of the initial temperature T0 0 041� . , 0.036, 0.031, 0.026, and 0.021 K (curves 1–5, respectively). It can be seen from Fig. 9 that the process of h-phonon creation is more rapid at higher values of the initial temperature. The dependence ? �( t shown in Fig. 9 is close to that obtained in [7,8], which was derived with the approximate cone distribution function (10), and with corresponding values of T p and " p . In Fig. 10,a the dependences E tl ( )� 0 (curve 1) and E tl ( )� 42 �s (curve 2) on the initial temperature T0, with ! 0 kept constant, are shown. We see, that with increasing of T0, the energy density E tl ( )� 42 �s also increases. In Fig. 10,b similar dependences of the angular widths $l t( )� 0 (curve 1) and $ �l t( )� 42 s (curve 2) are shown. All the curves shown on Fig. 10 were calculated with the value of ! 0 0 02� . which is a typical measured value. From Fig. 10 we see that with increasing initial tempera- ture, we expect wider and higher energy pulses on the detec- tor. Such behavior was observed in experiments when the initial power was increased, see Figs. 3 and 7 in Ref. 10. Creation of high-energy phonons by four-phonon processes in anisotropic phonon systems of He II7 Fizika Nizkikh Temperatur, 2007, v. 33, No. 5 533 a 1 2 3 4 0 5 10 15 20 25 30 35 40 b 4 3 2 1 0 5 10 15 20 25 30 35 40 T , K t, s� t, s� 0.060 0.056 0.052 0.048 0.044 0.06 0.05 0.04 0.03 0.02 Fig. 7. a — The temperature T t( ) is shown as a function of time, calculated with T0 0041� . K and ! 0 002� . (1), 0.025 (2), 0.03 (3), 0.035 (4). b — The anisotropy parameter ! �(t is shown as a function of time, calculated with ! 0 002� . and T0 0025� . (1), 0.030 (2), 0.036 (3), 0.041 (4) K. 8 7 6 5 4 t, s� 0 5 10 15 20 25 30 35 40 $ , d eg re e l 17 16 15 14 13 12 E , J/ m 3 l Fig. 8. The l-phonon energy density E tl( ) and the angular width of the l-phonons in momentum space $l t( ) are shown as functions of propagation time, calculated with initial values of ! 0 002� . , T0 0041� . K. 1 2 3 4 5 0.6 0.5 0.4 0.3 0.2 0.1 t, s� 0 5 10 15 20 25 30 35 40 Fig. 9. The energy density lost by the l-phonons, due to h-phonon creation, relative to the initial l-phonon energy den- sity, ? �(t , is shown as a function of time calculated with ! 0 002� . for different values of T0 equal to 0.041 (1), 0.036 (2), 0.031 (3), 0.026 (4), and 0.021 (5) K. 6. The momentum which separates the l- and h-phonon subsystems The calculations made above are based on the fact that phonons in superfluid helium separate into two subsystems according to their relaxation rate: l-phonon subsystem in which equilibrium occurs quickly and h-phonon subsystem in which equilibrium occurs slowly. The basis for such di- vision is the strong inequality (2). In this paper, as well as in previous ones (see Refs. 7, 8), it was supposed, that the momentum pd separating l- and h-phonon subsystems is equal to pc , after which the spontaneous decay of phonons is forbidden. However, according to Refs. 1, 19 the conser- vation laws of energy and momentum allow three-phonon processes ( )1 2 only up to momentum p pc1 2 4 5 � / (at the saturated vapour pressure ~p1 2 � 8.94 K with our parametrization of the dispersion curve). In the momentum range from p1 2 to pc processes of one phonon decaying into more than two phonons are allowed. So, for example, one phonon decay into three is allowed by conservation laws up to momentum p pc1 3 9 10 � / ( ~p1 3 � 9.49 K), into four — up to 9.7 K, into five — up to 9.81 K and so on up to ~pc � 10.0 K. At present there are no publications of calculations of the rates of the mentioned above processes, although 1 to 3 is submitted [20]. Hence there is the question of which subsystem phonons with momenta p p pc1 2 � � (82) should be assigned. If the rates of one phonon decay into three, four, etc. appear much higher than the rate �1 then these phonons should be part of the l-phonon subsystem, and so p pd c� . Otherwise these phonons should be part of the h-phonon subsystem and in this case p pd c� . The answer to this question is important as the numeri- cal value of the rate �1, for phonons moving at small an- gles to the anisotropy axis, is very sensitive to the value of pd when p pd c� . Such sensitivity is because the main contribution in Eq. (43) for �1 is given by phonons with momentum p3 close to the upper limit of integration pd . This reason for this is now explained. The distribution function n 2 0( ) exponentially decreases with increasing " 2. As a result, the main contribution to the integration over " 2 is given by phonons with " "2 12� min (see (61)). Hence the integration over " 2 leads to the following factor in the integrand of (43) exp expmin� � � �� � � �� = � � � �� � � �� cp k T c k TB B 2 12" . , (83) which for values of T expected in experiments, quickly increases with decreasing .. It is not difficult to check that ., determined by Eq. (30), monotonically decreases with increasing p3 and is minimal (and the exponent function is accordingly maxi- mal) when p pd3 � . So if T � 0 041. K, p pd c� then for ~p1 � 10.001 K and the typical value of ~p2 � 1.58 K when ~p3 � 10 K, the exponent of the exponential function � � � ; �c k TB. / .5 35 10 3, and when ~p3 � 8.95 K, we have � � �c k TB. / .4 67; i.e., the value of the exponential func- tion increases by two orders of magnitude. We see that re- 534 Fizika Nizkikh Temperatur, 2007, v. 33, No. 5 I.N. Adamenko, Yu.A. Kitsenko, K.E. Nemchenko, V.A. Slipko, and A.F.G. Wyatt 1 2 3 cp /k , K1 B 10 8 10 7 10 6 10 5 10 4 10 3 10 2 10 1 8.5 9.0 9.5 10.0 10.5 11.0 11.5 � 1 , s– 1 Fig. 11. The momentum dependence of the rate �1 at $1 0� for different values of ~pd equal to 8.94 (1), 9.49 (2), and 10 (3) K. These dependences were obtained from Eq. (43) with T � 0041. K and ! � 002. . a 0T , K 0.020 0.025 0.030 0.035 0.040 1 2 8 6 4 2 0 b l$ , d eg re e 2 1 18 16 14 12 10 0.020 0.025 0.030 0.035 0.040 0T , K l E , J/ m 3 Fig. 10. a — The initial l-phonon energy density, E tl ( )� 0 (curve 1), and the energy density after 42 �s, E tl ( )� 42�s (curve 2), are shown as functions of the initial temperature T0. b — The initial angular width of the l-phonons in momentum space, $l t( )� 0 (curve 1), and after 42 �s, $ �l t s( )� 42 (curve 2), are shown as functions of the initial temperature T0. All calculations were made with ! 0 002� . . ducing pd , and correspondly reducing the top limit of in- tegration over p3, leads to an exponential decrease of the rate �1( )p1 , at small $1. In Fig. 11 the momentum dependence of the rate �1 1( )p is shown at three different values of ~pd equal to 8.94, 9.49, and 10 K (curves 1–3, respectively) at $1 0� . These dependences were obtained from expression (43) at values of T � 0 041. K and ! � 0 02. . From Fig. 11 one can see, that with decreasing ~pd from 10 to 9.5 K, the rate �1 1 10( ~ )p � K decreases an order of magnitude. The physical causes for such a strong decrease are the strong anisotropy of l-phonon subsystem and the conser- vation laws (3) which make process 2 2 forbidden for angles $ $12 12� min . Such a high sensitivity to the choice of pd is absent in isotropic phonon systems. So when ! �1, the calculation of the rate with a help of formula (43), the rate �1 in isotropic phonon system, only changes 1.5 times, when pd goes from pc to p1 2 . So in order to answer the question of the value of pd in superfluid helium, it is necessary to compare results of the theory presented in this paper with experimental data [21], where the shape and the arrival time of h-phonon signal cre- ated by l-phonon pulse, at the bolometer, were investigated. It follows from the Eq. (25), that the exponential factor in N b ( )sh , cuts off at large angles $1 (see Fig. 3). In connection with this the rate �1, which is in N b ( )sh (see Eq. (25)) is of great interest especially at small values of $1. We note that in the case of small$1 at different values of pd , the rate �1 is described by curves similar to those in Fig. 11 for $1 0� . From Fig. 11 we see that changing pd , shifts the curves �1 1( )p to lower momenta by amounts equal to the differ- ences in the values of pd . Therefore if p pd c� , the maxi- mum h-phonon signal should be formed by phonons with momentum ~p1 = 10.0 K and phonons with ~p1 � 8.95 K should be absent. When p pd � 1 2 the maximum h-phonon signal should be due to phonons with ~p1 =8.95 K and phon- ons with momentum ~p1 �10 K should be practically absent. The group velocity vgr of phonons depends on momen- tum. From the data in Ref. 22, at the saturated vapour pressure, v pgr K( ~ . )� �8 95 207 m/s and v pgr K( ~ )� �10 = 189 m/s. Therefore the position of the maximum of the h-phonon signal, which in experiments [6,21] is measured with good accuracy, allows us to understand which phonons form this maximum and accordingly the correct value of pd . Experiments [6,21] show that the maximum of the h-phonon signal moves with a speed of 186.4 m/s, which corresponds to a group velocity of phonons with ~p1 � 10.15 K. Calculations similar to those presented in the previous section, with p pd c� , give the shape of h-phonon signal close to that observed in Refs. 6, 21, and are distinctly dif- ferent from the shape with smaller values of pd . From this section we see that in superfluid helium, the momentum pd separating the l- and h-phonon subsystems is equal to ~ .pc �10 0 K, as it was supposed in this paper and earlier ones, Refs. 7, 8. 7. Conclusion The general expression (38) for the rate �1 which de- scribes the decay (22) and creation (24) of high-energy phonons in a pulse of low-energy phonons moving from a heater to the detector, was found. Also the exact analytical expression for the rate �1 when $1 0� (see (43)) was found. This expression also describes the relaxation in isotropic phonon systems, i.e., when ! �1. From Eq. (43) we derive a simple and explicit analytical expression for the rate �1 for the case when h-phonons move along the anisotropy axis ("1 0� ) (see (57)). This allowed us to get an analytical approximation for �1 in the most interesting range of angles "1 (see (62)). Starting from Eq. (38), the rates were a numerically calculated. This allowed us to obtain the momentum and angular dependences of the rate �1 (see Fig. 1). Also the dependences of the rate �1 on the parameters of l-phonon pulse were obtained (see Fig. 2). It was shown, that angu- lar dependence of the rate �1 has a maximum at ~p1 13� K (see Fig. 1,a). The physical reasons of this maximum were given and the detailed analysis of such behavior of the rate �1 was made. The calculation of the rate �1 allowed us to get the an- gular distribution of created h-phonons and show that they occupy a solid angle, in momentum space, which is much smaller than the solid angle for l-phonons (see Figs. 4 and 5). This result explains the observation that the created h-phonon are concentrated near the anisotropy axis of the system [5]. Starting from the kinetic equation (73), the problem of the evolution of a short l-phonon pulse, caused by the cre- ation of h-phonons, was solved. For this, the principal consideration was the angular dependence of the collision integral N b ( )sh in (73), which quickly decreases with in- creasing angle (see Fig. 3). The physical reasons of such behavior were analyzed, and it was shown that it is con- nected with the exponential factor in N b ( )sh (see Eq. (25)). We investigated how the h-phonon creation rate de- pended on the parameters of the l-phonon pulse. It was shown that h-phonon creation not only increased the ani- sotropy parameter ! of l-phonon pulse but also increased the temperature T (see Fig. 7). However the energy den- sity of l-phonon pulse decreased as it must (see Fig. 8). For the conditions used in the experiments [13,14], the theory showed that up to 50% of initial energy of l-phonon pulse could be transferred into h-phonon creation (see Fig. 9). Also the dependence of the energy El and the angular width $l of the l-phonon pulse at detector, on the initial temperature T0, were investigated (see Fig. 10). It was shown, that with increasing initial temperature there will Creation of high-energy phonons by four-phonon processes in anisotropic phonon systems of He II7 Fizika Nizkikh Temperatur, 2007, v. 33, No. 5 535 be wider and higher energy density pulses at the detector. Such phenomenon were observed in experiments [10] when different heater powers were used. It was shown that the rate �1, for small angles $1, strongly depends on momentum pd which separates the l- and h-phonon subsystems (see Fig. 11). From a compari- son of the present theory with experimental data we see that p pd c� . Acknowledgments We express our gratitude to EPSRC of the UK (grant EP/C 523199/1) for support of this work. Appendix A. Expressions for the matrix elements Making the dot products in (17), (18), (20) and taking into account the conservation laws of energy and momen- tum (3) we have M u p p p p ( ) ( ) 1 1 2 1 2 1 2 12 12 1 2 12 1 2 2 2 2 1 2 � � � � � � � � � � � � � � � " " " � � � � � � � � � � � � � � � � � � � � � � � � � 2 2 1 2 34 34 3 4 34 3 4 2 u p p p p " " " ( ) � � � � � � � (A.1) M u p p p p ( ) ( ) 5 1 2 1 2 1 2 12 12 1 2 12 1 2 2 2 2 1 2 � � � � � � � � � � � � � � � " " " � � � � � � � � � � � � � � � � � � � � � � � � � 2 2 1 2 34 34 3 4 34 3 4 2 u p p p p " " " ( ) � � � � � � � � , (A.2) M u p p p p 13 3 1 3 1 3 1 3 13 13 1 3 13 1 3 2 2 1 2 ( ) ( ) � � � � � � � � � � � � � � " " " 2 24 24 2 4 24 2 4 2 2 2 1 2 � � � � � � � � � � � � � � � � � � � � � � � � u p p p p " " " ( ) � � � � � � � � � , (A.3) M u p p p p 24 3 1 3 1 3 1 3 13 13 1 3 13 1 3 2 2 1 2 ( ) ( � � � � � � � � � � � � � � � " " " ) ( )2 24 24 2 4 24 2 4 2 2 2 1 2 � � � � � � � � � � � � � � � � � � � � � � � u p p p p " " " � � � � � � � � � � , (A.4) M u p p p p 14 3 1 4 1 4 1 4 14 14 1 4 14 1 4 2 2 1 2 ( ) ( ) � � � � � � � � � � � � � � " " " 2 23 23 2 3 23 2 3 2 2 2 1 2 � � � � � � � � � � � � � � � � � � � � � � � � u p p p p " " " ( ) � � � � � � � � � , (A.5) M u p p p p 23 3 1 4 1 4 1 4 14 14 1 4 14 1 4 2 2 1 2 ( ) ( � � � � � � � � � � � � � � � " " " ) ( )2 23 23 2 3 23 2 3 2 2 2 1 2 � � � � � � � � � � � � � � � � � � � � � � � u p p p p " " " � � � � � � � � � � , (A.6) where " " 34 3 4 2 1 2 2 1 2 12 3 4 2 2 � � � � �( ) ( ) , p p p p p p p p (A.7) " " 23 1 4 2 2 3 2 1 4 14 2 3 2 2 � � � � �( ) ( ) , p p p p p p p p (A.8) 536 Fizika Nizkikh Temperatur, 2007, v. 33, No. 5 I.N. Adamenko, Yu.A. Kitsenko, K.E. Nemchenko, V.A. Slipko, and A.F.G. Wyatt " " 24 1 3 2 2 4 2 1 3 13 2 4 2 2 � � � � �( ) ( ) , p p p p p p p p (A.9) " " " " " " " " " , ,1 1 1 1 1 2 2 12 2i i i i i i� � � � � � �cos ( ), (A.10) Appendix B. Solution of combined equations (77) and (78) To solve combined equations (77) and (78) it is conve- nient to write them as � � � � � � � � : � : dE dt E d dt E cP E cP l h l l h h � , ( ) ( � � ). (B.1) The system of equations (B.1) can be solved for � (! �t and � ( )T t . This system of two equations with initial conditions ! � !( t � �0 0 and T t T( )� �0 0 can be solved numerically and the functions ! �( t and T t( ) found. However it is always convenient to carry out calculations with simple analytical expressions, which approximate the dependences of the functions in (94) on ! and T . Such expressions can be intro- duced as products of power functions and exponents. The form of these functions was obtained from (68), (70), (79), (80). The exponents of the power function and the coeffi- cients in exponential functions, were calculated by a least squares fitting. For the intervals corresponding to the conditions in ex- periments, 0 01 0 06. .� �! and 0 016 0 06. .K K� �T , ap- proximations of the functions contained in (B.1) can be written in SI system of units as E T Tl T T( , ) . . . . .! ! ! � � � �3713 36 0 503 0 0135 1 107 3 25e , (B.2) E T cP Tl l T T( , ) ( , ) . . . .! ! ! ! � � � � �445 856 0 435 0 00112 1 0 785e T 3 24. , (B.3) � ( , ) . . . . .E T Th T T! ! ! � ; � � �8 51 1013 1135 0 00648 1 0 831 4 75e , (B.4) � ( , ) � ( , ) . . . .E T cP Th h T T! ! ! ! � � ; � � �2 59 1012 1118 0 0158 1 1e 15 5 96T . . (B.5) Having substituted (B.2)–(B.5) into (B.1) and made the differentiation, we obtain the system of two equations with respect to �! and �T . The numerical solution of this sys- tem gives ! ! �( , ,t T0 0 and T t T( , , )! 0 0 . 1. I.N. Adamenko, Yu.A. Kitsenko, K.E. Nemchenko, V.A. Slipko, and A.F.G. Wyatt, Low Temp. Phys. 31, 459 (2005) [Fiz. Nizk. Temp. 31, 607 (2005)]. 2. I.N. Adamenko, Yu.A. Kitsenko, K.E. Nemchenko, V.A. Slipko, and A.F. G. Wyatt, Phys. Rev. B72, 054507 (2005). 3. I.N. Adamenko, K.E. Nemchenko, V.A. Slipko, and A.F.G. Wyatt, J. Phys.: Condens. Matter 17, 1 (2005). 4. A.F.G. Wyatt, N.A. Lockerbie, and R.A. Sherlock, J. Phys.: Condens. Matter 1, 3507 (1989). 5. M.A.H. Tucker and A.F.G. Wyatt, J. Phys.: Condens. Matter 6, 2813 (1994). 6. M.A.H. Tucker and A.F.G. Wyatt, J. Low Temp. Phys. 113, 621 (1998). 7. I.N. Adamenko, K.E. Nemchenko, A.V. Zhukov, M.A.H. Tucker, and A.F.G. Wyatt, Phys. Rev. Lett. 82, 1482 (1999). 8. A.F.G. Wyatt, M.A.H. Tucker, I.N. Adamenko, K.E. Nem- chenko, and A.V. Zhukov, Phys. Rev. B62, 9402 (2000). 9. I.N. Adamenko, K.E. Nemchenko, and A.F.G. Wyatt, J. Low Temp. Phys. 125, 1 (2001). 10. R. Vovk, C.D.H. Williams, and A.F.G. Wyatt, Phys. Rev. B68, 134508 (2003). 11. I.N. Adamenko, Yu.A. Kitsenko, K.E. Nemchenko, V.A. Slipko, and A.F.G. Wyatt, Phys. Rev. B73, 134505 (2006). 12. I.N. Adamenko, K.E. Nemchenko, and A.F.G. Wyatt, J. Low Temp. Phys. 126, 1471 (2002). 13. D.H.S. Smith, R. Vovk, C.D.H. Williams, and A.F.G. Wyatt, Phys. Rev. B72, 054506 (2005). 14. R. Vovk, C.D.H. Williams, and A.F.G. Wyatt, Phys. Rev. Lett. 91, 235302 (2003). 15. I.M. Khalatnikov, An Introduction to the Theory of Superfluidity, Addison-Wesley, Redwood City, CA (1989). 16. M.A.H. Tucker and A.F.G. Wyatt, J. Phys.: Condens. Matter 4, 7745 (1992). 17. V.M. Apalkov and M.E. Portnoi, Phys. Rev. B65, 125310 (2002). 18. V.M. Apalkov and M.E. Portnoi, Phys. Rev. B66, 121303 (2002). 19. J. Jackle and K.W. Kehr, Phys. Rev. Lett. 27, 654 (1971). 20. I.N. Adamenko, Yu.A. Kitsenko, K.E. Nemchenko, V.A. Slipko, and A.F.G. Wyatt, The Process of One Phonon Decaying into Three Phonons in Superfluid Helium, sub- mitted. 21. R.V. Vovk, C.D.H. Williams, and A.F.G.Wyatt, Phys. Rev. B69, 144524 (2004). 22. W.G. Stirling, in 75 th Jubilee Conference on Liquid He- lium-4, J.G.M. Armitage (ed.), World Scientic, Singapore (1983), p. 109. Creation of high-energy phonons by four-phonon processes in anisotropic phonon systems of He II7 Fizika Nizkikh Temperatur, 2007, v. 33, No. 5 537
id nasplib_isofts_kiev_ua-123456789-127809
institution Digital Library of Periodicals of National Academy of Sciences of Ukraine
issn 0132-6414
language English
last_indexed 2025-12-07T18:30:39Z
publishDate 2007
publisher Фізико-технічний інститут низьких температур ім. Б.І. Вєркіна НАН України
record_format dspace
spelling Adamenko, I.N.
Kitsenko, Yu.A.
Nemchenko, K.E.
Slipko, V.A.
Wyatt, A.F.G.
2017-12-28T12:54:17Z
2017-12-28T12:54:17Z
2007
Creation of high-energy phonons by four-phonon processes in anisotropic phonon systems of He II / I.N. Adamenko, Yu.A. Kitsenko, K.E. Nemchenko, V.A. Slipko, A.F.G. Wyatt // Физика низких температур. — 2007. — Т. 33, № 5. — С. 523-537. — Бібліогр.: 22 назв. — англ.
0132-6414
PACS: 67.70.+n, 68.08.–p, 62.60.+v
https://nasplib.isofts.kiev.ua/handle/123456789/127809
The problem of the creation of high-energy phonons (h-phonons) by a pulse of low-energy phonons&#xd; (l-phonons) moving from a heater to a detector in superfluid helium, is solved. The rate of h-phonon creation&#xd; is obtained and it is shown that created h-phonons occupy a much smaller solid angle in momentum space,&#xd; than the l-phonons. An analytical expression for the creation rate of h-phonon, along the symmetry axis of a&#xd; pulse, are derived. It allows us to get useful approximate analytical expressions for the creation rate of&#xd; h-phonons. The time dependences of the parameters which describe the l-phonon pulse are obtained. This&#xd; shows that half of the initial energy of l-phonon pulse can be transferred into h-phonons. The results of the&#xd; calculations are compared with experimental data and we show that this theory explains a number of experimental&#xd; results. The value of the momentum, which separates the l- and h-phonon subsystems, is found.
We express our gratitude to EPSRC of the UK (grant&#xd; EP/C 523199/1) for support of this work.
en
Фізико-технічний інститут низьких температур ім. Б.І. Вєркіна НАН України
Физика низких температур
Квантовые жидкости и квантовые кpисталлы
Creation of high-energy phonons by four-phonon processes in anisotropic phonon systems of He II
Article
published earlier
spellingShingle Creation of high-energy phonons by four-phonon processes in anisotropic phonon systems of He II
Adamenko, I.N.
Kitsenko, Yu.A.
Nemchenko, K.E.
Slipko, V.A.
Wyatt, A.F.G.
Квантовые жидкости и квантовые кpисталлы
title Creation of high-energy phonons by four-phonon processes in anisotropic phonon systems of He II
title_full Creation of high-energy phonons by four-phonon processes in anisotropic phonon systems of He II
title_fullStr Creation of high-energy phonons by four-phonon processes in anisotropic phonon systems of He II
title_full_unstemmed Creation of high-energy phonons by four-phonon processes in anisotropic phonon systems of He II
title_short Creation of high-energy phonons by four-phonon processes in anisotropic phonon systems of He II
title_sort creation of high-energy phonons by four-phonon processes in anisotropic phonon systems of he ii
topic Квантовые жидкости и квантовые кpисталлы
topic_facet Квантовые жидкости и квантовые кpисталлы
url https://nasplib.isofts.kiev.ua/handle/123456789/127809
work_keys_str_mv AT adamenkoin creationofhighenergyphononsbyfourphononprocessesinanisotropicphononsystemsofheii
AT kitsenkoyua creationofhighenergyphononsbyfourphononprocessesinanisotropicphononsystemsofheii
AT nemchenkoke creationofhighenergyphononsbyfourphononprocessesinanisotropicphononsystemsofheii
AT slipkova creationofhighenergyphononsbyfourphononprocessesinanisotropicphononsystemsofheii
AT wyattafg creationofhighenergyphononsbyfourphononprocessesinanisotropicphononsystemsofheii