Professional Documents
Culture Documents
Efficiency at Maximum Power For Classical Particle Transport
Efficiency at Maximum Power For Classical Particle Transport
Katja Lindenberg
Department of Chemistry and Biochemistry and BioCircuits Institute, University of California San Diego,
9500 Gilman Drive, La Jolla, California 92093-0340, USA
(Received 9 August 2012; published 24 October 2012)
We derive the explicit analytic expression for efficiency at maximum power in a simple model of classical
particle transport.
The average particle occupation while in contact with reservoir [cf. Eqs. (8) and (14)]. The above Poissonian steady state
, n() , is given by Eq. (14) does not, in general, obey detailed balance with
respect to the separate rates W () , and implies the following
n() = ex , (9)
nonzero entropy production:
where we have introduced the dimensionless quantity x ,
() ()
()
Wn,n1 st
pn1
x = ( ),
() ()
(10) Si = kB Wn,n1 st
pn1 Wn1,n pnst ln ()
n Wn1,n pnst
and = 1/kB T (kB = Boltzmanns constant). One can
() ()
() ()
()
rewrite this equation in the more familiar form () = + k+
= kB k+ k n ln ()
kB T () ln n() . k n
The requirement that pst = peq,() for W = W () leads to () ()
the following genuine condition of detailed balance for these = JN XN 0, (19)
rates with respect to its equilibrium distribution:
() eq,()
() where we have introduced the thermodynamic forces:
Wn,n1 pn1 = Wn1,n pneq,() . (11)
n()
In the following we will adopt the standard choice of transition XN() = kB ln . (20)
n
rates [7] that satisfy this condition (cf. law of mass action),
namely, Of particular interest to us is the situation in which a heat
current from a hot to a cold reservoir is used to drive particles
() () () ()
Wn+1,n = k+ Wn1,n = n k , (12) uphill from low to high chemical potential. To investigate this
with the n-independent rates obeying the balance condition case in more detail, we henceforth focus on the case of only
two reservoirs = 1,2, with reservoir 1 the hot reservoir and
() () ()
k+ = k n . (13) 2 the cold one, T (1) T (2) . Note that we have not taken the
thermal energy of the particles into account as this would
require the consideration of a third thermal reservoir, making
III. THERMAL ENGINE the comparison with Carnot efficiency more involved. At the
Having identified the thermodynamically correct expres- steady state, one finds, using Eqs. (9) and (13), the following
sions Eqs. (12) and (13) for the transition probabilities, we can explicit results for the fluxes:
proceed to a stochastic thermodynamic analysis (see [8] for a JN(1) = JN(2) = (ex1 ex2 ), (21)
brief review) of a system coupled to several reservoirs, with
different temperatures and chemical potentials. One easily JE(1) = JE(2) = JN(1) , (22)
verifies that the steady state solution of Eq. (4) is again a
Poisson distribution: JQ(1) = kB T (1) x1 JN(1) , (23)
nn
pnst = en , (14)
n! where we have introduced the rate
with a steady state average number of particles reflecting the (1) (2)
k k
influence of each reservoir [compare with Eq. (13)]: = (1) (2)
. (24)
k + k
()
k
n = + ()
. (15) The entropy production reduces to the simple expression
k Si = kB (x2 x1 )(ex1 ex2 ) 0. (25)
This is most easily demonstrated
by showing that the generat-
ing function is given by n s n pnst = e(s1)n The power P of the engine, being the amount of net
Even though Poissonian, this distribution corresponds to a chemical energy produced per unit time, is given by
nonequilibrium steady state. Using Eqs. (12)(15), we find the P = ((2) (1) )JN(1)
following explicit expressions for the separate particle, energy,
and heat fluxes from each reservoir into the system: = kB T (1) [x1 (1 c )x2 ] (ex1 ex2 ). (26)
()
JN() = st
Wn,n1 pn1 ()
Wn1,n pnst = k+() ()
k n, (16) The corresponding efficiency reads
n ((2) (1) )JN(1) x2
JE() = JN() , (17) = = 1 (1 c ) . (27)
JQ(1) x1
JQ() = JE() () JN() = ( () )JN() . (18) Before turning to the main issue of efficiency at maximum
power, we first note that equilibrium, i.e., zero entropy
Note that the fluxes from each reservoir to the system are production [cf. Eq. (25)], is attained when x1 = x2 . This does
strongly coupled; i.e., energy, heat, and particle flux are not require that the temperatures T (1) and T (2) and chemical
proportional to each other. This is a result of our assumption potentials (1) and (2) be separately equal, a feature which is
that all particles have the same energy . Furthermore, the well known for strongly coupled systems [9]. In the vicinity
fluxes from reservoir are only zero when n = n() , i.e., when of such a point, the machine can operate reversibly, and its
the steady state distribution is the equilibrium distribution efficiency attains Carnot efficiency, = C [cf. Eq. (27)].
041144-2
EFFICIENCY AT MAXIMUM POWER FOR CLASSICAL . . . PHYSICAL REVIEW E 86, 041144 (2012)
Let us now turn our attention to the point of maximum where n represents the number of particles of species X, which
power. From can transmute into the species X() whose chemical potential
is fixed by reservoir constraints. This representation is most
P P
= = 0, (28) natural in an isothermal system, T () = T .
x1 x2 Finally, we mention a similarity of the model studied here
one finds with that of an underdamped Brownian particle in contact with
ln(1 C ) several heat baths. The Langevin equation for such a particle
x1 = 1 (1 C ) , (29) has the following form:
C
ln(1 C )
x2 = 1 . (30) mv = { () v + () T () () }, (33)
C
[1] F. Curzon and B. Ahlborn, Am. J. Phys. 43, 22 (1975); [3] M. Esposito, K. Lindenberg, and C. Van den Broeck, Phys. Rev.
A. Gomez-Marin and J. M. Sancho, Phys. Rev. E 74, 062102 Lett. 102, 130602 (2009).
(2006); C. Van den Broeck, Adv. Chem. Phys. 135, 189 (2007); [4] C. Van den Broeck, J. Stat. Mech. (2010) P10009.
A. E. Allahverdyan, R. S. Johal, and G. Mahler, Phys. Rev. [5] M. Esposito, K. Lindenberg, and C. Van den Broeck, Europhys.
E 77, 041118 (2008); T. Schmiedl and U. Seifert, Europhys. Lett. 85, 60010 (2009); M. Esposito, R. Kawai, K. Lindenberg,
Lett. 81, 20003 (2008); Y. Izumida and K. Okuda, ibid. 83, and C. Van den Broeck, Phys. Rev. E 81, 041106 (2010).
60003 (2008); Phys. Rev. E 80, 021121 (2009); M. Esposito, [6] See, for example, G. H. Wannier, Statistical Physics (Dover,
R. Kawai, K. Lindenberg, and C. Van den Broeck, Phys. Rev. New York, 1987).
Lett. 105, 150603 (2010); Y. Zhou and D. Segal, Phys. Rev. E 82, [7] N. G. Van Kampen, Stochastic Processes in Physics and
011120 (2010); U. Seifert, Phys. Rev. Lett. 106, 020601 (2011); Chemistry (North-Holland, Amsterdam, 1981).
N. Golubeva, A. Imparato, and L. Peliti, Europhys. Lett. 97, [8] C. Van den Broeck (unpublished).
60005 (2012); C. Van den Broeck, N. Kumar, and K. Lindenberg, [9] C. Van den Broeck, Adv. Chem. Phys. 135, 189 (2007).
Phys. Rev. Lett. 108, 210602 (2012); Y. Apertet, H. Ouerdane, [10] Z. C. Tu, J. Phys. A 41, 312003 (2008).
C. Goupil, and Ph. Lecoeur, Phys. Rev. E 85, 041144 (2012); [11] See, for example, J. M. R. Parrondo and P. Espagnol, Am. J.
Y. Wang and Z. C. Tu, ibid. 85, 011127 (2012). Phys. 64, 1125 (1996); P. Visco, J. Stat. Mech. (2006) P06006;
[2] C. Van den Broeck, Phys. Rev. Lett. 95, 190602 C. Van den Broeck and M. Esposito, Phys. Rev. E 82, 011144
(2005). (2010).
041144-3