fourier’s heat conduction equation

Upload: larissa-villela

Post on 02-Mar-2018

230 views

Category:

Documents


0 download

TRANSCRIPT

  • 7/26/2019 Fouriers Heat Conduction Equation

    1/22

    FOURIERS HEAT CONDUCTION EQUATION:

    HISTORY, INFLUENCE, AND CONNECTIONS

    T. N. NarasimhanDepartment of Materials Science and Mineral

    EngineeringDepartment of Environmental Science, Policy, and

    Management

    Lawrence Berkeley National LaboratoryUniversity of California, Berkeley

    Abstract. The equation describing the conduction ofheat in solids has, over the past two centuries, proved tobe a powerful tool for analyzing the dynamic motion ofheat as well as for solving an enormous array of diffu-sion-type problems in physical sciences, biological sci-ences, earth sciences, and social sciences. This equationwas formulated at the beginning of the nineteenth cen-tury by one of the most gifted scholars of modern sci-ence, Joseph Fourier of France. A study of the historicalcontext in which Fourier made his remarkable contribu-tion and the subsequent impact his work has had on thedevelopment of modern science is as fascinating as it is

    educational. This paper is an attempt to present a pic-ture of how certain ideas initially led to Fouriers devel-opment of the heat equation and how, subsequently,Fouriers work directly influenced and inspired others touse the heat diffusion model to describe other dynamicphysical systems. Conversely, others concerned with thestudy of random processes found that the equationsgoverning such random processes reduced, in the limit,to Fouriers equation of heat diffusion. In the process ofdeveloping the flow of ideas, the paper also presents, tothe extent possible, an account of the history and per-sonalities involved.

    1. INTRODUCTION

    The equation describing the conduction of heat insolids occupies a unique position in modern mathemat-ical physics. In addition to lying at the core of theanalysis of problems involving the transfer of heat inphysical systems, the conceptual-mathematical structureof the heat conduction equation (also known as the heatdiffusion equation) has inspired the mathematical for-mulation of many other physical processes in terms ofdiffusion. As a consequence, the mathematics of diffu-sion has helped the transfer of knowledge relating toproblem solving among diverse, seemingly unconnecteddisciplines. The transient process of heat conduction,described by a partial differential equation, was first for-mulated by Jean Baptiste Joseph Fourier (17681830)and presented as a manuscript to the Institut de Francein 1807. At that time, thermodynamics, potential theory,and the theory of differential equations were all in the

    initial stages of their formulation. Combining remark-able gifts in pure mathematics and insights into obser-vational physics, Fourier opened up new areas of inves-tigation in mathematical physics with his 1807masterpiece,Theorie de la Propagation de la Chaleur dans

    les Solides.Fouriers work was subjected to review by some of the

    most distinguished scientists of the time and was notaccepted as readily as one might have expected. It wouldbe another 15 years before this major contributionwould be accessible to the general scientific communitythrough publication of his classic monograph, Theorie

    Analytique de la Chaleur(Analytic Theory of Heat) [Fou-rier, 1822]. Soon after this publication, the power andsignificance of Fouriers work was recognized outside ofFrance. Fouriers method began to be applied to analyzeproblems in many fields besides heat transfer: electricity,chemical diffusion, fluids in porous media, genetics,and economics. It also inspired a great deal of re-search into the theory of differential equations.Nearly 2 centuries later, the heat conduction equationcontinues to constitute the conceptual foundation onwhich rests the analysis of many physical, biological, andsocial systems.

    A study of the conditions that led to the articulationof the heat conduction equation and the reasons why

    that equation has had such a major influence on scien-tific thought over nearly 2 centuries is in itself instruc-tive. At the same time, an examination of how the workwas received and accepted by Fouriers peers and suc-cessors gives us a glimpse into the culture of science,especially during the nineteenth century in Europe. Thepresent work has been motivated both by the educa-

    Copyright 1999 by the American Geophysical Union. Reviews of Geophysics, 37, 1 / February 1999

    pages 151172

    8755-1209/99/1998RG900006$15.00 Paper number 1998RG900006

    151

  • 7/26/2019 Fouriers Heat Conduction Equation

    2/22

    tional and historical importance of Fouriers work. Ac-cordingly, the purpose of this paper is to explore how theframework of the heat conduction equation has come tohelp us understand an impressive array of seeminglyunconnected natural processes and, in so doing, to gainhistorical insights into the manner in which scientificideas develop. The focus of this paper is on the connec-tions of concepts and ideas. An in-depth treatment ofFouriers 1807 manuscript can be found in the work ofGrattan-Guinness and Ravetz [1972], and critical treat-ment of Fourier the man and the physicist, includingimportant Fourier correspondence, can be found in thework ofHerivel[1975].

    The paper starts with scientific developments duringthe eighteenth century that set the stage for Fourierswork on heat conduction. Following this, details arepresented of Fourier himself and his contribution, espe-cially the 1807 manuscript. Fouriers influence has oc-curred along two lines. Experimentalists in electricity,chemical diffusion, and fluid flow in porous materials

    directly derived benefit from interpreting their experi-ments by analogy with the heat conduction phenome-non. Researchers in other fields such as statistical me-chanics and probability theory indirectly establishedconnections with the heat conduction equation by rec-ognizing the similarities between the mathematical be-havior of their systems and mathematical solutions ofthe heat conduction equation. These direct and indirectinfluences of Fouriers work are described next. Thepaper concludes with some reflections on the scientificatmosphere during the nineteenth century, a compari-son of the different facets of diffusion, and a look be-yond Fouriers solution strategy. A chronology of theimportant developments is presented in Table 1.

    2. DEVELOPMENTS LEADING UP TO FOURIER

    Before we describe the scientific developments of theeighteenth century that set the stage for Fouriers con-

    TABLE 1. Chronology of Significant Contributions on Diffusion

    Year Contribution

    Fahrenheit 1724 mercury thermometer and standardized temperature scaleAbbe Nollet 1752 observation of osmosis across animal membraneBernoulli 1752 use of trigonometric series for solving differential equationBlack 1760 recognition of latent heat and specific heatCrawford 1779 correlation between respiration of animals and their body heatLavoisier and Laplace 1783 first calorimeter; measurement of heat capacity, latent heat

    Laplace 1789 formulation of Laplace operatorBiot 1804 heat conduction among discontinuous bodiesFourier 1807 partial differential equation for heat conduction in solidsFourier 1822 Theorie Analytique de la ChaleurOhm 1827 law governing current flow in electrical conductorsDutrochet 1827 discovery of endosmosis and exosmosisGreen 1828 formal definition of a potentialGraham 1833 law governing diffusion of gasesThomson 1842 similarities between equations of heat diffusion and electrostaticsPoiseuille 1846 experimental studies on water flow through capillariesGraham 1850 experimental studies on diffusion in liquidsFick 1855 Fouriers model applied to diffusion in liquidsDarcy 1856 law governing flow of water in porous mediaDupuit 1863 potential theory applied to flow in groundwater basinsMaxwell 1867 diffusion equation for gases derived from dynamical theory

    Pfeffer 1877 investigations on osmosis in biological and inorganic membranesEdgeworth 1883 law of error and Fourier equationForchheimer 1886 flow nets for solving seepage problems using potential theory

    vant Hoff 1887 theory of osmotic pressure by analogy with gas lawsNernst 1888 interpretation of Ficks law in terms of forces and resistancesLord Rayleigh 1894 random mixing of sound waves as a diffusion processRoberts-Austen 1896 experimental measurement of solid diffusionBachelier 1900 option pricing and diffusion of probabilityEinstein 1905 Brownian motion and diffusion equationPearson 1905 notion of random walkBuckingham 1907 diffusion of multiple fluid phases in soilsLangevin 1908 framework for stochastic differential equationGardner 1922 measurement of potential in a multiple-fluid-phase porous mediumTerzaghi 1924 seepage in deformable clays as analogous to heat diffusionRichards 1931 nonlinear diffusion of moisture in soilsFermi 1936 neutron diffusion in graphite as analogous to heat diffusionBullard 1949 thermal gradient probe for the ocean floorTaylor 1953 advective dispersion as a diffusion process

    152 Narasimhan: FOURIERS EQUATION 37, 1 / REVIEWS OF GEOPHYSICS

  • 7/26/2019 Fouriers Heat Conduction Equation

    3/22

    tribution, it is useful to briefly state the nature andcontent of the heat conduction process. The transientheat diffusion equation pertains to the conductive trans-port and storage of heat in a solid body. The body itself,of finite shape and size, communicates with the externalworld by exchanging heat across its boundary. Within thesolid body, heat manifests itself in the form of temper-ature, which can be measured accurately. Under these

    conditions, Fouriers differential equation mathemati-cally describes the rate at which temperature is changingat any location in the interior of the solid as a functionof time. Physically, the equation describes the conserva-tion of heat energy per unit volume over an infinitesi-mally small volume of the solid centered at the point ofinterest. Crucial to such conservation of heat is therecognition that heat continuously moves across thesurfaces bounding the infinitesimal element as dictatedby the variation of temperature from place to placewithin the solid and that the change in temperature at apoint reflects the change in the quantity of heat stored inthe vicinity of the point.

    It is clear from the above that the notions of temper-ature, quantity of heat, and transport of heat, as well asthe relation between quantity of heat and temperature,are fundamental to Fouriers heat conduction model. Itis important to recognize here that these basic notionswere still evolving when Fourier developed his equation.Therefore it is appropriate to begin by familiarizingourselves with the evolution of these notions during theeighteenth century.

    Since heat can be readily observed and measured onlyin terms of temperature, the development of a reliablethermometer capable of giving repeatable measure-

    ments was critical to the growth of the science of heat.Gabriel Daniel Fahrenheit (16861736), a German in-strument maker and physicist, perfected the closed-tubemercury thermometer in 1714 and was producing itcommercially by 1717 [Middleton, 1966]. By 1724 he hadestablished what we now know as the Fahrenheit scalewith the melting of ice at 32and the boiling of water at212.

    The next developments of interest were qualitativeand conceptual, and of great importance. Joseph Black(17281799), a pioneer in quantitative chemistry, wasknown for his lectures in chemistry at Glasgow and wasalso a practicing physician. Around 1760 he noticed thatwhen ice melts, it takes in heat without changing tem-perature. This observation led him to propose the termlatent heat to denote the heat taken up by water as itchanges its state from solid to liquid. He also noticedthat equal masses of different substances needed differ-ent amounts of heat to raise their temperatures by thesame amount. He coined the term specific heat todenote this type of heat. Although Black is said to haveconstructed an ice calorimeter, he never published hisresults. The precise measurement of latent heat andspecific heat was left to Lavoisier and Laplace, some 20years later. Another important development was the

    appearance of the bookExperiments and Observations ofAnimal Heat, and the Inflammation of Combustible Bod-iesby Adair Crawford (17481795) in 1779. In this work,Crawford proposed that oxygen was involved in thegeneration of heat by animals during respiration andwent on to discuss a method of measuring specific heatby a method of mixtures [Guerlac, 1982]. Crawfordsidea of measuring specific heat by the method of mix-

    tures would soon have a significant influence onLavoisier and Laplace, although he himself was unableto measure these quantities accurately.

    In the wake of the contributions of Black and Craw-ford, what must be considered as one of the most im-portant papers of modern chemistry and thermodynam-ics appeared in 1783. This was the paper entitled

    Memoire sur la Chaleurcoauthored by Antoine LaurentLavoisier (17431794), the central figure of the revolu-tion in chemistry of the latter half of eighteenth century,and Pierre Simon Laplace (17491827), one of the moreinfluential mathematicians and theoretical physicists ofmodern science. Lavoisier and Laplace [1783] provided

    detailed descriptions of an ice calorimeter with whichthey measured, for the first time, the latent heat ofmelting of ice and the specific heats of different mate-rials. All the measurements were made relative to water,the chosen reference. They also showed experimentallythat animals release heat during respiration by placing aguinea pig within the calorimeter for several hours andmeasuring the quantity of ice melted. In a related set ofexperiments, they also demonstrated quantitatively thatthe process of respiration, in which oxygen is combinedwith carbon in the animals body, is in fact combustion,resulting in the release of heat. During the late nine-

    teenth century, when this work was done, the nature ofheat was still a matter of debate. Some believed that heatwas a fluid diffused within the body (referred to ascaloric) while others believed that heat was a mani-festation of vibrations or motions of matter at the atomiclevel. Although Lavoisier and Laplace preferred thelatter concept, they interpreted and presented their re-sults in such a way that the experiments stood by them-selves, independent of any hypothesis concerning thenature of heat. The significance of the Lavoisier-Laplacecontribution to Fouriers equation is that it provided thenotion of specific heat, which is fundamental to theunderstanding of time-dependent changes of tempera-ture. Nonetheless, the significance of the work far tran-scends Fouriers equation. By experimentally quantify-ing latent heat and heats of reactions, the Lavoisier-Laplace work constitutes an essential component of thefoundations of thermodynamics.

    We now consider the process of transfer of heat insolids, that is, the process of heat conduction. The bestknown pre-Fourier work in this regard is that of JeanBaptiste Biot (17741862) who made important contri-butions in magnetism, optics, and celestial mechanics.

    Biot[1804] addressed the problem of heat conduction ina thin bar heated at one end. In the bar, heat not only

    37, 1 / REVIEWS OF GEOPHYSICS Narasimhan: FOURIERS EQUATION 153

  • 7/26/2019 Fouriers Heat Conduction Equation

    4/22

    was conducted along the length but was also lost to the

    exterior atmosphere transverse to the direction of con-duction. Biots starting point was Newtons law of cool-

    ing, according to which the rate at which a body losesheat to its surroundings is proportional to the difference

    in temperature between the bar and the exterior atmo-

    sphere. Biot, who was a student of Laplaces mechanisticschool, believed in the philosophy of action at a distance

    between bodies. Accordingly, the temperature at a pointin the heated rod was perceived to be influenced by all

    the points in its vicinity. Essentially, then, the mathemat-

    ical problem of heat conduction came to be consideredas one of a class of many-body problems. As was pointed

    out byGrattan-Guinness and Ravetz[1972], Biots ideal-ization of action at a distance involved only the differ-

    ence in temperature between points and did not involve

    the distance between the points. Consequently, Biotsapproach did not involve a temperature gradient, so

    necessary to the formulation of the differential equation.However, Biot did articulate the underlying concepts

    clearly by stating that when the heat content of the barchanges at each instant, the net accumulation of heat at

    a point causes a change in temperature. Biot also as-

    serted that he experimentally found Newtons law con-cerning the loss of heat to be rigorous. Similar inferences

    of a qualitative nature had been drawn by Lambert[1779], who had experimentally studied heat conduction

    in a rod [Herivel, 1975].

    Apart from these foundational developments relatingto heat, developments relating to potential theory and

    differential equations during the nineteenth century de-serve notice. The theory of potentials arises in many

    branches of science, such as electrostatics, magnetostat-ics, and fluid mechanics. Potential theory involves prob-lems describable in terms of a partial differential equa-

    tion in which the dependent variable is the appropriatepotential (defined as a quantity whose gradient is force)

    and the sum of the second spatial derivatives of thepotential in three principal directions is equal to zero.

    This equation was first formulated by Laplace [1789],

    although the term potential would be coined later byGeorge Green (17931841), a self-educated mathemati-cian, [Green, 1828].Laplace[1789] formulated the equa-tion in the context of the problem of the stability of

    Saturns rings.The eighteenth century also saw very active develop-

    ments in the theory of ordinary and partial differentialequations through the contributions of Daniel Bernoulli(17001782), Jean le Rond dAlembert (17171783),Leonhard Euler (17071783), John-Louis Lagrange(17361813), and others. For the partial differentialequation describing a vibrating string, Bernoulli hadsuggested, on physical grounds, a solution in terms oftrigonometric series. Similar usage of trigonometric se-ries was also made a little later by Euler and Lagrange.Yet dAlembert, Euler, and Lagrange were not particu-

    larly satisfied with the trigonometric series. Their con-

    cerns were purely mathematical in nature, consisting ofissues of convergence and algebraic periodicity of such

    series [Grattan-Guinness and Ravetz, 1972; Herivel,1975].

    In nineteenth-century Europe, two philosophical

    views of the physical world prevailed: the mechanisticschool of Isaac Newton (16421727) and the dynamic

    school of Gottfried Wilhelm Leibniz (16461716). Dur-ing the eighteenth and nineteenth centuries, a number of

    leading thinkers from France were fully committed to

    the mechanistic view and devoted their efforts to de-scribing the physical world with greater detail in terms of

    Newtons laws. At the same time, Newtons contempo-rary Leibniz also had a major influence on the develop-

    ment of scientific thought. At the foundation of physics

    were the notions of force, momentum, work, and action.Although these notions are all related, Newton and

    Leibniz pursued two parallel but distinct avenues tounderstanding the physical world. Newtons approach

    was based on the premise that by knowing forces andmomenta at every point or particle, one could com-

    pletely describe a physical system. Leibniz, on the other

    hand, pursued the approach of understanding the totalsystem in terms of work and action. One of the leading

    figures of Newtons mechanistic school was Laplace.Laplace, in turn, had many ardent followers, including

    Biot and Poisson. Among those who followed Leibnizs

    philosophy were Lagrange, Euler, and Hamilton. Al-though both approaches ultimately proved equivalent,

    the mathematics associated with each of them are verydifferent. While the mechanistic school relied on the use

    of vector fields to describe the physical system, thedynamic school of Leibniz, remarkably, realized thesame results through the use of energy and action, which

    are scalar quantities. Additionally, the thinking of themathematical physicists of the late eighteenth century

    was also influenced by their intense interest in celestialmechanics, a field that had greatly captivated Galileo,

    Newton, and Kepler.

    It was under these circumstances that observationaldata on heat, electricity, chemical reactions, and physi-ology of animals were being collected and great effortswere being made to understand them rationally in terms

    of force, momentum, energy, and work. As was alreadynoted, at the turn of the nineteenth century, the natureof heat was still unresolved. Those of the mechanisticschool, including Biot, believed that heat was a perme-ating fluid. On the other hand, those of the dynamicschool believed that heat was essentially motion, consist-ing of rapid molecular vibrations. Those of the mecha-nistic school also believed that a cogent theory of heatshould be rigorously built from a detailed description ofmotion at the level of individual particles. This ap-proach, it appears, influenced the work of Biot [1804]and his use of action at a distance.

    154 Narasimhan: FOURIERS EQUATION 37, 1 / REVIEWS OF GEOPHYSICS

  • 7/26/2019 Fouriers Heat Conduction Equation

    5/22

    3. FOURIERS CONTRIBUTION

    The science of heat, the theory of potentials, and thetheory of differential equations were all in their earlystages of development by the time Fourier started hiswork on heat conduction. Opinions were still dividedabout the nature of heat. However, heat conduction dueto temperature differences and heat storage and the

    associated specific heat of materials had been experi-mentally established. Potential theory had already beenformulated. Finally, the representation of dynamic prob-lems in continuous media with the help of partial differ-ential equations (e.g., the problem of a vibrating string)and their solution with the help of trigonometric serieswere also known. In this setting, Fourier began workingon the transient heat conduction problem.

    Fouriers life and contributions are so unusual that abrief sketch of his career and the conditions under whichhe worked is worthwhile. For a comprehensive account,the reader is referred to Grattan-Guinness and Ravetz[1972] and Herivel [1975]. Fourier was born in 1768 in

    Auxerre in Burgundy, now the capital of Yonne depart-ment in central France. In 1789, about the time hismathematical talents began to blossom, the French Rev-olution intervened. In his native Auxerre he was sociallyand politically active, being a forceful orator. His out-spoken criticism of corruption almost took him to theguillotine in 1794; he was saved mainly by the publicoutcry in the town and a deputation of local people onhis behalf. Following this he taught mathematics for afew years at the Ecole Polytechnique in Paris. In 1798,Napoleon Bonaparte (17691821) was leading an expe-dition to Egypt, and Fourier was made Secretaire Per-

    petuel of the newly formed Institut dEgypte. In Egypthe held many important administrative and judicial po-sitions and, in 1799, was made leader of a scientificexpedition investigating monuments and inscriptions inUpper Egypt. In November 1801, Fourier returned toFrance upon the withdrawal of French forces fromEgypt. However, his hopes for resuming his teachingduties at the Ecole Polytechnique were ended whenNapoleon made him prefect of the department of Isere,near the Italian border, with its capital at Grenoble.

    During his tenure as prefect, Fourier embarked ontwo very different major scholarly efforts. On the onehand, he played a leadership role on a multivolume workon Egypt, which would later form the foundation for thescience of Egyptology. On the other, he began workingon the problem of heat diffusion. It appears that Fourierstarted work on heat conduction sometime between1802 and 1804, probably for no other reason than that hesaw it as one of the unsolved problems of his time.Between 1802 and 1807 he conducted his researches intoEgyptology and heat diffusion whenever he could findspare time from his prefectural duties.

    Like Biot before him, Fourier initially formulatedheat conduction as an n-body problem, stemming fromthe Laplacian philosophy of action at a distance. During

    these early investigations he was aware of Biots work,having received a copy of the paper from Biot himself.For some reason that is not quite clear, Fourier aban-doned the action at a distance approach around 1804and made a bold departure from convention, whicheventually led to his masterpiece, the transient heatconduction equation.

    Essentially, Fourier moved away from discontinuous

    bodies and towards continuous bodies. Instead of start-ing with the basic equations of action at a distance,Fourier took an empirical, observational approach toidealize how matter behaved macroscopically. In thisway he also avoided discussion of the nature of heat.Rather than assuming that the behavior of temperatureat a point was influenced by all points in its vicinity,Fourier assumed that the temperature in an infinitesimallamina or element was dependent only on the conditionsat the lamina or element immediately upstream anddownstream of it. He thus formulated the heat diffusionproblem in a continuum.

    In formulating heat conduction in terms of a partial

    differential equation and developing the methods forsolving the equation, Fourier initiated many innovations.He visualized the problem in terms of three compo-nents: heat transport in space, heat storage within asmall element of the solid, and boundary conditions. Thedifferential equation itself pertained only to the interiorof the flow domain. The interaction of the interior withthe exterior across the boundary was handled in terms ofboundary conditions, conditions assumed to be knowna priori. The parabolic equation devised by Fourier wasa linear equation in which the parameters, conductivity,and capacitance were independent of time or tempera-

    ture. This attribute of linearity enabled Fourier to drawupon the powerful concept of superposition to combinemany particular solutions and thereby create generalsolutions [Grattan-Guinness and Ravetz, 1972]. The su-perposition artifice offered such promise for solvingproblems that mathematicians who followed Fourierresorted to linearizing differential equations so as tofacilitate their subsequent solution.

    Perhaps the most powerful and most daunting aspectof Fouriers work was the method of solution. Fourierwas clearly aware of the earlier work of Bernoulli, Euler,and Lagrange relating to solutions in the form of trigo-nometric series. He was also aware that Euler,DAlembert, and Lagrange viewed trigonometric serieswith great suspicion. Their opposition to the trigonomet-ric series stemmed from reasons of pure mathematics:convergence and algebraic periodicity. Lagrange, in fact,had a particular preference for solutions expressed inthe form of Taylor series [Grattan-Guinness and Ravetz,1972]. Yet Fourier, who was addressing a well-definedphysical problem with physically realistic solutions, didnot allow himself to be held back by the concerns of hisillustrious predecessors. He boldly applied the methodof separation of variables and generated solutions interms of infinite trigonometric series. Later, he would

    37, 1 / REVIEWS OF GEOPHYSICS Narasimhan: FOURIERS EQUATION 155

  • 7/26/2019 Fouriers Heat Conduction Equation

    6/22

    also generate solutions in the form of integrals thatwould come to be known as Fourier integrals. In the lastpart of his 1807 work, Fourier also presented someresults pertaining to heat conduction in a cylindricalannulus, a sphere, and a cube.

    Fourier submitted his manuscript to the FrenchAcademy in December 1807. As was the practice, thesecretary of the Academy appointed a committee of

    reviewers consisting of four of the most renowned math-ematicians of the time, Laplace, Lagrange, Monge, andLacroix. The manuscript was not well received, particu-larly by Laplace and Lagrange, for the mathematicalreasons alluded to above. Although Laplace would laterbecome sympathetic to Fouriers method, Lagrangewould never change his mind. Because of the lack ofapproval by his peers, the possible publication of Fouri-ers work by the French Academy was delayed indefi-nitely. In the end, Fourier took it upon himself to expandthe work and publish it on his own in 1822 under the titleTheorie Analytique de la Chaleur; it is now an avowedclassic.

    4. THE HEAT CONDUCTION EQUATION

    It is appropriate to introduce here the transient heatconduction equation of Fourier. In modern notation,this parabolic partial differential equation may be writ-ten as,

    KT cT

    t , (1)

    whereK is thermal conductivity, T is temperature, c isspecific heat capacity of the solid per unit volume, and tis time. The dependent variable Tis a scalar potential,while thermal conductivity and specific heat capacity areempirical parameters. Physically, the equation expressesthe conservation of heat per unit volume over an infin-itesimally small volume lying in the interior of the flowdomain. The exchange of heat with the external world isto be taken into account with the help of either temper-ature or thermal fluxes prescribed on the boundary.Also, it is assumed that the distribution of temperatureover the domain is known at the initial time t 0. Forthe particular case when the temperature over the flowdomain does not change with time and Kis independentof temperature, (1) reduces to Laplaces equation.

    In (1), thermal conductivity Kis physically a constantof proportionality, which relates the quantity of heatcrossing a unit surface area in unit time to the spatialgradient of temperature perpendicular to the surface.This relationship is now known as Fouriers law. In his1807 manuscript, Fourier formulated thermal conductiv-ity mathematically rather than experimentally. As waspointed out by Grattan-Guinness and Ravetz [1972] and

    Herivel[1975], Fourier arrived at this concept gradually,as he was making the transition from discontinuous

    bodies to a continuous body. The notion of heat flux wasyet a new concept, and Fourier would fully clarify it onlyin 1810, in a letter to an unknown correspondent [Heri-

    vel, 1975].The concept of specific heat capacity, proposed ex-

    perimentally by Lavoisier and Laplace [1783], is an es-sential part of the transient heat diffusion process. Ithelps convert the rate at which heat is accumulating in

    an elemental volume to an equivalent change in temper-ature. Thermal conductivity and thermal capacity aretwo different attributes of a solid, one governing trans-port in space and the other governing change in storagein the vicinity of a point. Together, these two parametersgovern the ability of the solid to respond in time toforces that cause the thermal state of the solid to change.Sometimes, it is found mathematically convenient tocombine the two parameters into a single parameterknown as thermal diffusivity, K/(c), where isdensity of the solid. The higher the diffusivity, the fasterthe tendency of the material to respond to externallyimposed perturbations.

    5. INFLUENCE AND CONNECTIONS

    Soon after the publication of the analytic theory ofheat in 1822, the general scientific community becameaware of the significance of Fouriers work, not merelyfor the science of heat, but in general as a rationalframework for conceptualization for other branches ofscience. Within a few years the heat conduction analogywas brought to the study of electricity, and later it wasapplied to the analysis of molecular diffusion in liquids

    and solids. The dynamical theory of gases directly led tothe analogy between diffusion of gases and diffusion ofheat. The investigation of the flow of blood throughcapillary veins and the flow of water through porousmaterials led to the adaption of Fouriers heat conduc-tion model to the flow of fluids in geologic media. Thestudy of random motions of particles led to the inter-pretation of Fouriers equation in terms of stochasticdifferential equations.

    Simultaneously, Fouriers work began also to be rec-ognized by the establishments of the intellectual world[Grattan-Guinness and Ravetz, 1972]. He was made aforeign member of the Royal Society in 1823, and in1827 he was elected to the Academie Francaise and theAcademie de Medicine. He succeeded Laplace as thepresident of the Council of Prefects of the Ecole Poly-technique. He also became the Secretaire Perpetuel ofthe Academie des Sciences.

    For the sake of completeness it may be mentionedhere that Fouriers political career came to an end withthe fall of Napoleon at Waterloo in 1815. His pensionwas refused and, close to 50 years old, he was virtuallywithout an income. However, thanks to a former studentof his at the Ecole Polytechnique in 1794 who was aprefect of the department of Seine, Fourier was given

    156 Narasimhan: FOURIERS EQUATION 37, 1 / REVIEWS OF GEOPHYSICS

  • 7/26/2019 Fouriers Heat Conduction Equation

    7/22

    the directorship of the Bureau of Statistics in Paris.Later, in 1817, he was elected to a vacancy in physics inthe Academie des Sciences. With these appointments,Fourier had a secure income for the rest of his life andhe could find plenty of time for conducting research.During the 1820s Fourier also had an influential anddistinguished following: Sturm, Navier, Sophie Germain,Dirichlet, and Liouville.

    To gain an understanding of Fouriers influence overthe past nearly 2 centuries, it is convenient to organizethe discussions into the following general subheadings:electricity, molecular diffusion, flow in porous materials,and stochastic diffusion.

    5.1. ElectricityThe nature of electricity and its relation to magnetism

    were not completely understood at the time Fourierpublished his analytic theory, nor were the relationsbetween electrostatics and electrodynamics (galvanicelectricity). Quantities such as current strength and in-tensity were not precisely defined. At this time, Georg

    Simon Ohm of Germany (17871854) set himself thetask of removing the ambiguities about galvanic electric-ity with mathematical rigor, supported by experimentaldata. He published four papers on galvanic current be-tween 1825 and 1827, of which the most well-known ishis 1827 pamphlet, Die galvanische Kette, mathematisch

    Bearbeitet. Ohms work, which is considered to be one ofthe most important fundamental contributions to elec-tricity, was largely inspired by Fouriers heat conductionmodel.Ohm[1827] started with three laws. Accordingto his first law, the communication of electricity fromone particle takes place only directly to the particle next

    to it, so that no immediate transition from that particleto any other situated at a greater distance occurs. Recallthat Fourier made this important idealization whenmaking the transition from action at a distance to thecontinuous medium. The second law was that of Cou-lomb, relating to the effect of a charge at a distance in adielectric medium. The third law was that when dissim-ilar bodies touch one another, they constantly maintainthe same difference of potential at the surface of con-tact. This assumption is quite important because itpoints to a significant difference between the processesof heat conduction and conduction of electricity. In thecase of heat conduction, temperature is continuous atmaterial interfaces, whereas in the case of galvanic elec-tricity the potential, namely, voltage, is discontinuous, asis implied by this assumption of Ohm.

    Ohms careful experiments showed that the current ina galvanic circuit did not vary with time (steady flow), theintensity of the electric current (measured with a torsionmagnetometer) was directly proportional to the drop inelectrostatic force (measured with an electroscope)along the conductor in the direction of flow and in-versely proportional to the resistance of the conductor.In turn, the resistance of the conductor was a function ofthe material of which the conductor is made and of its

    form (resistance was found to be inversely proportionalto the cross-sectional area of the conductor). Equallyimportant, Ohm showed that the resistance of the con-ductor was independent of the magnitude of the currentitself or the magnitude of the electrostatic force. Ohmgave a precise meaning to flux (current) and resistance.However, he erred in his use of electrostatic force, whichhe considered to be the quantity of electricity contained

    in an unit volume. Although electrostatic force so de-fined is an intensive quantity, it was left to Kirchhoff[1849] to establish that Ohms law should properly beexpressed in terms of potential (voltage) differencerather than difference in electrostatic force [Archibald,1988]. Following Ohms work, the measurement of theelectrical resistance of various materials with great pre-cision became a fundamental task in physics [Maxwell,1881].

    Ohm took the analogy with heat conduction fartherand considered the flow of electricity to be exactly anal-ogous to the flow of heat and wrote a transient equationof the form similar to (1),

    du

    dt

    d2u

    dx 2

    bc

    u, (2)

    where is a quantity analogous to heat capacity,u is theelectrostatic force, is electrical conductivity, b is atransfer coefficient associated with the atmosphere towhich electricity is being lost by the conductor accordingto Coulombs law,c is the circumference of the conduc-tor, and is the area of cross section of the conductoralong the x direction. (Unless otherwise stated, the no-tations used in this paper are those of the referenced

    authors.) Ohm was not confident about this equationand admitted that no experimental evidence forwas asyet forthcoming.

    James Clerk Maxwell (18311879) derived the sameequation in a different context and showed that Ohmwas in error in proposing (2) the way he did. Maxwell[1881] considered a long conducting wire (such as atransoceanic telegraph cable) surrounded by an insula-tor. In this case, the insulator, which is a dielectricmaterial, functions as a condenser and possesses theelectrical capacitance property analogous to heat capac-itance. Moreover, if the insulator is not perfect, someamount of electricity would be lost to the surroundings,as is indicated by the second term on the right-hand sideof (2). Maxwell [1881, p. 422] expressed Ohms errorthus: Ohm, misled by the analogy between electricityand heat, entertained an opinion that a body whenraised to a high potential becomes electrified throughoutits substance, as if electricity were compressed into it,and was thus by means of an erroneous opinion led toemploy the equations of Fourier to express the true lawsof conduction of electricity through a long wire, longbefore the real reason of the appropriateness of theseequations had been suspected. Indeed, it is fundamen-tal to the nature of electricity that capacitance is an

    37, 1 / REVIEWS OF GEOPHYSICS Narasimhan: FOURIERS EQUATION 157

  • 7/26/2019 Fouriers Heat Conduction Equation

    8/22

    electrostatic phenomenon and only insulators possessthat property. Electricity, as Maxwell [1881, p. 336]pointed out, behaves like an incompressible fluid, andhence conductors do not possess the property of capac-itance.

    It is interesting to note that Ohm formulated his fluxlaw in terms of a difference in potential and a resistance,rather than in terms of the infinitesimal notion of a

    gradient as was done by Fourier. The resistance, inOhms law is an integral that combines the materialproperty as well as the geometry of the conductor offinite size through which current is flowing. Fouriersmethod of separating material property from geometrywas of the right mathematical form to pose the problemas a differential equation. In contrast, Ohms approachof dealing with resistance and potential difference ismore naturally suited for appreciating the diffusionproblem directly in terms of integrals involving work,energy, and action.

    Ohms work is now accepted as one of the mostimportant contributions in the science of electricity. Yet

    recognition did not come to him readily. Although phys-icists such as Fechner, Lenz, Weber, Gauss, and Jacobidrew upon Ohms work in their own research soon afterOhm publishedDie galvanische Kette, Ohms work cameunder criticism from an unexpected quarter. His exper-imental approach to finding order in nature was heavilycriticized by Georg Poul [Gillispie, 1981], a physicist whowas a follower of Hegels philosophy of pure reason.However, due recognition came to Ohm after a fewyears when he was elected to the Academies at Berlinand Munich and the Royal Society conferred on him theCopley Medal in 1841.

    William Thomson (18241907), also known as LordKelvin, was greatly influenced by Fouriers work. Thom-sons first two articles, written at ages 16 and 17, were indefense of Fouriers mathematical approach. Later, hedemonstrated the similarities between the mathematicalstructures of Fouriers heat conduction equation and theequations of electrostatics stemming from the works ofLaplace and Poisson [Thomson, 1842]. For example,potential was analogous to temperature, a tube of induc-tion was analogous to a tube of heat flow, the electro-motive force was in the direction of the gradient ofpotential and the flux of heat was in the direction oftemperature gradient.

    While physical analogies serve a useful purpose,Max-well [1888, pp. 5253] emphasized that caution was inorder to prevent the analogies from being carried toofar. He pointed out that the analogy with electric phe-nomena applied only to the steady flow of heat. Evenhere, differences exist between electricity and heat. Forsteady flow, heat must be kept up by a continuoussupply, accompanied by its continuous loss. However, inelectrostatics a set of electrified bodies placed in a per-fectly insulating medium might remain electrified for-ever without any supply from external sources. More-over, there is nothing in the electrostatic system that can

    be described as flow. Note also that the temperature ofa body cannot be altered without altering the physicalstate of the body, such as density, conductivity, or elec-trical properties. On the contrary, bodies may bestrongly electrified without undergoing any physicalchange.

    It is pertinent here to mention a major geologicalcontroversy of the nineteenth century in which Lord

    Kelvin and the heat conduction model played a part.Fourier himself had maintained on more than one oc-casion [Herivel, 1975] that the phenomenon of terrestrialheat motivated him to develop a theory for heat conduc-tion in solids. On the basis of geological observations,contemporary geologists were of the opinion that theEarth was very old. For example, Charles Darwin hadestimated that the age of the Earth was about 300million years, based on assumed erosional rates of sed-iments. Kelvin analyzed the problem from a differentbasis, assuming that the Earth was initially a solid sphereat a high uniform temperature which gradually lost heatby conduction to reach the present state. Accordingly, he

    estimated that the Earth cannot be older than about 100million years [Hallam, 1983]. Based on this he severelycriticized Darwin and other geologists for grossly over-estimating the age of the Earth. However, Kelvins cool-ing Earth model was eventually invalidated by with thediscovery of radiogenic heat in the Earths crust. Theradioactive heat source of the Earth enabled Earth sci-entists to extend the age of the Earth to many billionyears during the early twentieth century.

    Despite his erroneous estimate of the age of theEarth, Kelvins conceptualization of global heat trans-port was very perceptive. It drew the attention of Earth

    scientists to the fact that the Earth is a heat engine andthat observations of temperature and heat flow near theEarths surface are essential for understanding the in-ternal structure and the evolution of the Earth. Prior tothe 1950s, however, few heat flow measurements wereavailable from continental boreholes, and practicallynothing was known about the natural loss of heat fromthe oceanic floors occupying over 70% of the Earthssurface. The prevalent untested view was that becausethe oceanic crust was known to have very low content ofradioactive minerals (compared to the high content ofthe continental granitic rocks), heat loss from the floorsof the oceans must be significantly smaller than that ofthe continental surface.

    A major breakthrough in the field of terrestrial heatflow studies was the design and development of a probeto measure temperature gradients on the deep oceanfloor by Edward Crisp Bullard (19071980) in the sum-mer of 1949 [Bullard et al., 1956]. This device producedthe first set of heat flow data from the Pacific Ocean in1952 and the Atlantic Ocean in 1954. Surprisingly, thedata showed that the heat loss from beneath the oceanswas comparable in magnitude to that from the conti-nents, and it became necessary for Earth scientists torethink their fundamental notions about the thermal

    158 Narasimhan: FOURIERS EQUATION 37, 1 / REVIEWS OF GEOPHYSICS

  • 7/26/2019 Fouriers Heat Conduction Equation

    9/22

    structure of the Earth. These initial observations pro-vided a major impetus to marine geophysicists to themeasurement of heat flow beneath the deep oceans.

    This was also the period during which the notion ofplate tectonics was taking root in the Earth sciences onthe basis of diverse field observations: magnetic rever-sals, young ages of oceanic crust, patterns of submarineseismicity and volcanism. Within a decade of the first

    heat flow measurements by Bullard and his coworkers,by the mid 1960s, several hundred heat flow measure-ments became available from the Atlantic, the Pacificand the Indian Oceans. The spatial patterns of heat flowon the ocean floors revealed by these measurementscontributed greatly to the emerging notions of platetectonics. Fundamental to the new view was the hypoth-esis that the oceanic crust is created at the mid-oceanicridges by molten rock welling up from the Earths inte-rior and that upon cooling, the rigid oceanic platespreads away from the ridges with time. Thus the rocksof the seafloor become older with distance away fromthe ridge.

    To analyze this problem,McKenzie[1967] consideredthe oceanic crust to be a plate of finite thickness (about50 km), in which conductive heat transport occurred intwo dimensions. The bottom (550C) and the top of theplate (0C) were treated as constant temperature bound-aries. The lateral spread of the plate away from the ridgeat a finite velocity of a few centimeters per year washandled with the help of an advection (linear transla-tion) term. Heat flow observations from the Atlantic, thePacific and the Indian Oceans agreed reasonably wellwith the estimates based on the solutions of this advec-tion-diffusion problem. The fact that the spatial heat

    flow patterns from the ocean floors were consistent witha spreading seafloor with vertical heat conductionthrough the rigid crust was an important corroborativefactor in the establishment of plate tectonics as a viabletheory.

    We saw earlier that Ohm had attempted unsuccess-fully to formulate a time-dependent electrical flow equa-tion by direct analogy with Fouriers equation. Laterwork, stemming from Maxwells equations, establishedthat transient heat conduction and transient electricityflow are very different in nature. Transient flow of elec-tricity typically arises in the case of alternating current asopposed to the steady state direct current with whichOhm was concerned. In the case of alternating current,the change in electric field is intrinsically coupled withan induced magnetic field in a direction perpendicular tothe direction in which current is flowing. The nature ofthe coupled phenomena is such that when the frequencyof the alternating current is low, Maxwells electromag-netic equations may be described in the form of anequation which looks mathematically similar to the heatconduction equation, in that one side of the equationinvolves the Laplace operator (second derivative inspace) and the other involves the first derivative in time.However, the resemblance is only superficial because the

    dependent variable in this equation is a vector potential,whereas the dependent variable in the heat conductionequation is a scalar potential.

    5.2. Molecular DiffusionMolecular diffusion is the process by which molecules

    of matter migrate within solids, liquids and gases. Thephenomenon of diffusion was observationally known to

    chemists and biologists during the eighteenth century. Inthe early nineteenth century, experimental chemists be-gan paying serious attention to molecular diffusion, andthe publication of Fouriers work provided the chemistswith a logical framework with which to interpret andextend their experimental work. The following discus-sion on molecular diffusion starts with diffusion in liq-uids, followed by solids and gases.

    5.2.1. Diffusion in liquids. Among the earliestobservations that attracted the attention of chemists todiffusion in liquids is the phenomenon of osmosis. In1752, Jean Antoine (Abbe) Nollet (17001770) ob-served and reported selective movement of liquids

    across an animal bladder (semipermeable membrane).Between 1825 and 1827, Joachim Henri Rene Dutrochet(17761847) made pioneering contributions in the sys-tematic study of osmosis. A physiologist and medicaldoctor by training, Dutrochet spent most of his career inthe study of the physiology of animals and plants. Aboutthis time, Poisson had attempted to explain osmosis interms of capillary theory. Dutrochet [1827] strongly dis-agreed with Poisson and, on the basis of experimentalevidence, argued that two currents (solute and solvent)simultaneously occur in opposite directions during os-mosis, one of them being stronger than the other, and

    that the understanding of osmosis required somethingmore than a simple physical mechanism such as capil-larity. He speculated on the possible role of electricity inthe osmotic phenomenon. He also coined the termsendosmosis for the migration of the solvent toward thesolution and the term exosmosis for the reverse pro-cess.

    The next major work on liquid diffusion was that ofThomas Graham (18051869).Graham[1850] presenteddata on the diffusibility of a variety of solutes andsolvents in his Bakerian Lecture of the Royal Society.Despite the wealth of data he collected, Graham did notattempt to elicit from them a unifying fundamentalstatement of the process of diffusion in liquids. ThatGraham restricted himself essentially to the collection ofexperimental data on diffusion in liquids proved to be acatalyst for one of the most influential papers of molec-ular diffusion, that ofFick[1855a, b]. Adolf Fick (1829 1901) was a Demonstrator in Anatomy at the Universityof Zurich and, in addition to his professional training inmedicine, had a sound background in mathematics andphysics. Fick expressed regret that Graham failed toidentify any fundamental law of diffusion from his sub-stantial experimental data. In seeking to remedy thesituation, Fick saw a direct analogy between the diffu-

    37, 1 / REVIEWS OF GEOPHYSICS Narasimhan: FOURIERS EQUATION 159

  • 7/26/2019 Fouriers Heat Conduction Equation

    10/22

    sion of heat in solids and the diffusion of solutes inliquids.

    By direct analogy with Fourier, Fick [1855b] wrotedown the parabolic equation for transient diffusion ofsolutes in liquids in one dimension thus:

    D2c

    x2

    1

    Q

    dQ

    dx

    c

    x c

    t, (3)

    where D is the diffusion coefficient, c is aqueous con-centration, andQ is the area of cross section. Note thatFick made a novel departure from Fourier in writing theone-dimensional equation. The second term on the left-hand side of (3) accounts for the variation of the area ofcross section along the flow path (the x axis). Intrinsi-cally, Ficks equation is valid for a flow tube of arbitraryshape involving a curvilinear x axis. Indeed,Fick[1855a,b] presented data from a diffusion experiment in aninverted-funnel-shaped vessel, solved (3) for the geo-metric attributes of the particular cone-shaped vessel,and found that his mathematical solution compared

    favorably with the steady state concentrations at differ-ent locations within the vessel. For a flow tube withconstant area of cross section, (3) simplifies to Fouriersequation. One can readily verify that (3) leads also toappropriate differential equations for radial and spher-ical coordinates. Upon reflection, it becomes apparentthat integration of (3) along curvilinear flow tubes leadsto the evaluation of resistances within finite segments offlow tubes and that the evaluation of resistances thusprovides a link between the approaches of Fick andOhm.

    According to Fick, concentration is analogous to tem-

    perature, heat flux is analogous to solute flux and ther-mal diffusivity is analogous to chemical diffusivity. Ifconcentration in the aqueous phase is defined as massper unit volume, then specific chemical capacity (analo-gous to specific heat) equals unity and chemical diffu-sivity is equal to chemical conductivity.

    In the second part of his paper, Fick[1855a, b] wenton to analyze flow across a semipermeable membrane byidealizing it as a collection of cylindrical pores of radius. As was suggested earlier by Dutrochet [1827], twosimultaneous currents will occur through the capillaries;the solute current will occur toward the solvent and thesolvent current will occur toward the solution. Fick rea-soned that because of the affinity of water to the mate-rial composing the membrane, the water current will beorganized more toward the walls of the pores and thesolute will be organized toward the axis of the pores.Incidentally, a remarkably similar reasoning was em-ployed by Taylor[1953], who studied solute diffusion incapillary tubes with moving water. When the radius ofthe pore becomes sufficiently small, the flow of thesolute will be arrested and osmosis will involve onecurrent, that of the solvent.

    The study of liquid diffusion was soon to take a veryimportant place in the field of biophysics through the

    investigations of Wilhelm Pfeffer (18451920). Afterreceiving a doctoral degree in chemistry from the Uni-versity of Gottingen when he was 20 years old, Pfeffergrew interested in the study of biological processes andbrought his experimental and analytical skills to bear onthe study of mass transfer in plant cells. Broadly, treatingthe outer layer of the cell as a semipermeable mem-brane, Pfeffer devised sophisticated techniques to mea-

    sure osmotic pressure within cells and went on to de-velop and test several hypotheses concerning thediffusion of nutrients within and across cells.

    Pfeffer found osmosis experiments on plant cells to belimiting and sought to conduct measurements on con-trolled inorganic membranes. Along these lines he pio-neered the use of thin layers of ferrocyanide depositedon ceramic substrates as semipermeable membranes.Using such membranes he measured osmotic pressure ofvarious solutions as a function of concentration as wellas temperature. Pfeffers data, published in his 1877classic, Osmotische Untersuchungen, would later helpvant Hoff to lend credibility to his theory of osmotic

    pressure. Dutrochet, Pfeffer, Fick, and other biophysi-cists of the time strongly supported the view that phys-iological processes must be elucidated and understood interms of inorganic (nonbiological) processes.

    By the time Pfeffer published his book on cell me-chanics, a wealth of data had been collected on osmosis,both from physiological and inorganic materials. Manyhypotheses were in vogue, and a rational description ofosmosis in terms of known principles of physics andchemistry was lacking. Jacobus Hendricus vant Hoff(18521911), an influential physical chemist of the sec-ond half of the nineteenth century, filled this gap by

    providing a theoretical foundation for osmotic pressurebased on well-established laws of chemistry. Vant Hoff[1887] started with and justified the proposition that thephysical behavior of solutions and the associated os-motic pressure can be rationally understood by treatingsolutions as analogous to gases and by applying Boyleslaw, Gay-Lussacs law, and Avogadros law to solutions.He formally defined osmotic pressure as the excess pres-sure that would develop in a solution contained in avessel that communicates with a reservoir of a solventacross a perfect semipermeable membrane. By using theaforesaid laws and the second law of thermodynamics,vant Hoff was able to draw many inferences aboutrelationships between the magnitude of osmotic pres-sure on the one hand and the nature of the soluteconcentration and temperature on the other. He dem-onstrated that the experimental data of previous work-ers, especially Pfeffer [1877], substantially justified histheoretical framework.

    In osmosis, two opposing currents of flow are in-volved, each being driven by its own force: the solvent byspatial variations in its fluid potential, and the solute bythe spatial variations of osmotic pressure. Therefore it isconvenient to conceptualize the total pressure in thesolution as a sum of the water phase pressure and the

    160 Narasimhan: FOURIERS EQUATION 37, 1 / REVIEWS OF GEOPHYSICS

  • 7/26/2019 Fouriers Heat Conduction Equation

    11/22

    osmotic pressure. Thus in the solution, the pressure inthe water phase pw ptotal posmotic. The stronger theconcentration of the solution, the lesser the water phasepressure and the stronger will be the solvent currenttoward the solution should the solution communicatewith the solvent. Analogously, the solute will be driven inthe opposite direction because osmotic pressure de-creases in the direction of the solvent.

    Closely following vant Hoff, Walther HermannNernst (18641941) examined the process of solute dif-fusion in the context of osmotic pressure as defined bythe former.Nernst[1888] pointed out that the diffusionof solutes in the direction of decreasing concentrationhad been suggested earlier by Berthollet [1803] and thatFick established it rigorously with mathematics, sup-ported by experimental data. Nernst found Ficks ap-proach to be formal and lacking in the elucidation of theforces that impelled the solute diffusion process. Toovercome this deficiency, he looked at diffusion in termsof impelling forces and resistive forces, the former stem-ming from spatial variations of osmotic pressure and the

    latter stemming from the collision of molecules with thesolvent molecules and even among the solute moleculesthemselves.

    Nernst [1888] considered the force due to osmoticpressure acting on a molecule of the solute and defineda coefficient of resistance Krepresenting the force re-quired to move 1 gram-molecule of the solute throughthe solvent at a velocity of 1 cm s1. Combining these, heexpressed the flux of solute in terms of the gradient ofthe osmotic pressure and the reciprocal of the coefficient

    K. He then recognized that for dilute solutions, osmoticpressure is linearly related to concentration by a simple

    relation,p p0c, wherep0 is the osmotic pressure in asolution containing a gram molecular weight of the sol-ute and c is concentration. As a result, for dilute solu-tions, the ratiop0/Kbecomes part of the diffusion coef-ficient and flux becomes proportional to the gradient ofconcentration, as was proposed by Fick [1855a, b]. Byextending the analysis to concentrated solutions, Nernstpointed out that in such solutions the solute will encoun-ter greater resistance to flow because of mutual collisionamong the solute molecules in addition to the solventmolecules. Therefore in concentrated solutions the dif-fusion coefficient will be a function of concentration. Asa consequence, the relevant differential equation of dif-fusion becomes nonlinear.

    For electrolytes in which individual ions will migrateseparately, Nernst suggested that in order that the ionscomposing a given solute may migrate at the same ve-locity, the differences in ion velocities induced by os-motic pressure will be compensated by electrostaticforces.

    5.2.2. Solid diffusion. An early documented ob-servation of solid diffusion is attributed to Robert Boyle(16271691) [Barr, 1997], who succeeded in 1684 inmaking zinc diffuse into one of the faces of a copperfarthing, leading to the formation of brass. By carefully

    filing the face, Boyle showed that zinc had indeed dif-fused into the body of copper. Yet controlled diffusionmeasurements would not become possible until 200years later.

    The first measurements of the diffusion of one solidmetal into another was made by William Roberts-Aus-ten (18431902), who was Chemist and Assayer of theBritish Mint. He took up the challenge of extending

    Grahams work on liquid diffusion to metals. Hisprogress was considerably hampered by difficulties inaccurately measuring the temperature at which diffusionwas taking place in the solid state. By adopting LeChateliers platinum-based thermocouples, he suc-ceeded in studying the diffusion of gold in solid lead atdifferent temperatures. The results were analyzed interms of Fouriers model of one-dimensional diffusion[Roberts-Austen, 1896].

    Solid diffusion is a process of great importance inmany natural and industrial processes. In the lithosphereof the Earth it influences the genesis of minerals, ores,and rocks. As an illustration, we consider the role of

    solid diffusion in metamorphism, one of three majorrock-forming processes. Metamorphism constitutes thephysical, chemical, and structural adjustments under-gone by solid rocks in response to temperature, pressure,and other environmental changes consequent to burialat depths within the Earths crust. Metasomatism is ametamorphic process by which a new mineral may growin the body of an old mineral or mineral aggregate, oftenoccurring at constant volume, with little disturbance oforiginal texture or structure of the mineral (palimpsesttextures).

    Many geologists of the first half of this century, espe-

    cially from Europe, were persuaded by field evidence tobelieve that solid diffusion must have been responsiblefor the genesis of large bodies of granite and granite-looking rocks, which were previously thought to havesolidified from a molten magma. Although their infer-ence was justified qualitatively, the magnitude of the roleof solid diffusion in the genesis of large rock massescould not be gauged reliably. This was due to a lack ofinstruments to observe the effects of diffusion on themicroscopic scale at which solid diffusion occurs withinmineral grains. Fortunately, the 1950s saw rapid devel-opments in the field of X-ray spectrometry, leading tothe development of the electron-microprobe. This re-markable instrument enabled quantitative chemicalanalysis in situ within a mineral grain, in the immediatevicinity of a location. By the mid-1960s, mineralogistswere able to quantitatively study variations of chemicalcomposition within a single mineral grain at a spatialresolutions of1 m. New possibilities opened up in theapplication of the diffusion equation to address funda-mental questions related to the age of rocks and theirthermal history.

    Over the past 3 decades, diffusion-based mathemati-cal models have been used to decipher the history ofcertain potassium-rich minerals such as feldspars using

    37, 1 / REVIEWS OF GEOPHYSICS Narasimhan: FOURIERS EQUATION 161

  • 7/26/2019 Fouriers Heat Conduction Equation

    12/22

    the decay of radioactive 40K to 40Ar. Similarly, tech-niques have also been developed to analyze the decay ofuranium in naturally occurring zircon to lead. In bothcases the assumption is that the radioactive clock startsto count when the mineral of interest (feldspar or zir-con) begins to crystallize from a melt at a certain criticaltemperature. Once the crystal has formed, the decayprocess will release the daughter product, as dictated by

    the appropriate half-life. For different chemical reasons,the daughter products will tend to be excluded from thelattice structure and diffuse out of the crystal, at a ratethat depends on the geometry of the crystal (sheet,sphere, cylinder), the diffusion coefficient, and the ther-mal environment. In minerals, diffusion coefficient is anexponential function of temperature (the Arrhenius re-lationship), and diffusion will become practically negli-gible below some threshold temperature. Thus soliddiffusion in these minerals is constrained to a tempera-ture window. Consequently, the measurement of diffu-sion profiles within a single mineral grain may providevaluable information about the thermal history of the

    mineral and its environment, that is, whether it cooledrapidly or very slowly. In quantitatively interpreting ob-served diffusion profiles, it is customary to use solutionsto Fouriers heat diffusion equation for assumed simplegeometries of the mineral grains (sheet, sphere, cylin-der), initial conditions and boundary conditions.

    Although looking at minerals at extremely fine spatialresolution is of great value, rocks and their origin have tobe understood on a much larger field scale, on the scaleof rock masses. On a large scale, one may wish to look atmetasomatism as a macroscopic solid diffusion process.Intuitively, this may be reasonable because the porosity

    of rocks is extremely small, comprising microfissures andgrain boundaries. However, such a macroscale idealiza-tion may lead to quantitatively inaccurate inferencesbecause the fluids in the micropores and grain bound-aries may have diffusivities that are many orders ofmagnitude larger than the corresponding diffusivities ofthe solid minerals. Thus, in these larger bodies, diffusionover long distances may be dictated by the fluids in thefine pores, and solid diffusion may occur locally, withinthe mineral grains on the small scale. This entails theapplication of Fouriers diffusion equation to systemswith interpenetrating continua (e.g., fractures on the onehand, and the solid on the other) with radically differentdiffusion properties.

    We may end this discussion on solid diffusion with thework of Enrico Fermi (19011954) on neutron diffusion.Fermi was the first to successfully achieve, in 1942, asustained release of energy from a source other than theSun by bombarding and splitting uranium atoms with thehelp of neutrons slowed down in a matrix of solid graph-ite. Critical to the design of the experiment was thecalculation of the slowing down of neutrons and theabsorption of thermal neutrons by the carbon host. Theslowing down of the neutrons was described as a diffu-sion process [Anderson and Fermi, 1940], and the corre-

    sponding diffusion constants were calculated on the ba-sis of experimental data. The approach was one Fermihad already perfected earlier [Fermi, 1936]. The diffu-sion theory developed by Fermi would later be known asthe age theory.

    5.2.3. Diffusion of gases. The earliest experimen-tal work on the diffusion of gases was byGraham[1833].When two or more gases are mixed together in a closed

    vessel, the natural tendency is for the gases to redistrib-ute themselves by diffusion in such a way that the mix-ture has a uniform composition everywhere. Grahamshowed experimentally that the rate at which each of thegases diffuses is inversely proportional to the square rootof its density. This observation is known as Grahamslaw. When we compare gas diffusion with liquid diffu-sion or the conduction of heat or electricity, we find thatin these latter cases we are concerned with conductivetransport of the permeant in different host materials,whereas in this case of gas diffusion we are concernedwith the conduction of gas in free space. In the case ofnongaseous conduction, the transport coefficient (con-

    ductivity or diffusivity) is experimentally estimated fordifferent materials on the basis of Fouriers law. Thusconductivity is a property of the material rather than thepermeant. In contrast, in the case of pure gaseous dif-fusion, diffusivity is a property that stems solely from theattributes of the permeating fluid, the gas.

    With the advances that were taking place in molecu-lar physics and chemistry during the middle of the nine-teenth century, a great deal of effort was made byresearchers to directly estimate the properties of gasessuch as viscosity, specific heat, thermal conductivity,diffusion coefficient, and diffusivity by starting with

    force, momentum, and energy at the molecular level andstatistically integrating these quantities in space andtime to estimate the macroscopic properties of interest.Among the earliest researchers in this regard was Max-well whose work on the dynamical theory of gases is offundamental importance. Maxwell [1867] assumed mol-ecules to be small bodies or groups of small bodies whichpossess forces of mutual repulsion varying inversely asthe fifth power of distance. Macroscopically, he de-scribed the diffusion of a mixture containing two gases interms of an equation with the same form as Fourierstransient heat conduction equation. In this case, thediffusion coefficient is describable by Daltons law ofpartial pressures and densities of the two gases and isinversely proportional to the total pressure. Maxwellgenerated a solution of this equation for the case of aparticular column experiment conducted by Grahaminvolving carbonic acid and air and found some agree-ment with the diffusion coefficient independently esti-mated by Graham.

    5.3. Flow of Water in Porous MaterialsFouriers heat conduction equation has had an enor-

    mous influence in the study of fluid flow processes in theEarth, especially water and petroleum in porous media.

    162 Narasimhan: FOURIERS EQUATION 37, 1 / REVIEWS OF GEOPHYSICS

  • 7/26/2019 Fouriers Heat Conduction Equation

    13/22

    In applying the equation to these processes, the follow-ing analogies can be made: temperature corresponds toscalar fluid potential, heat corresponds to mass of fluid,thermal conductivity corresponds to hydraulic conduc-tivity and, heat capacity corresponds to hydraulic capac-ity. However, unlike electricity, heat, and solutes, thepotential of water has a very special attribute, namely,gravity. This attribute renders the extension of heat

    analogy to the Earth sciences particularly interesting.5.3.1. Steady flow of water. By the time of Fou-riers work, fluid mechanics was a well-developed sci-ence, and the concept of a fluid potential, defined asenergy per unit mass of water was already establishedthrough the seminal contribution of Bernoulli in hydro-dynamics, early in the eighteenth century. The flow ofwater in open channels was being rigorously studied bycivil engineers. In addition to civil engineers, many phys-iologists were also interested in the study of water flowthrough capillary tubes to better understand, by analogy,the flow of blood through narrow vessels.

    Among the earliest experimentalists to study the slow

    motion of water through capillary tubes was Jean Leo-nard Marie Poiseuille (17991869), a physician andphysiologist. Not satisfied with the contemporary under-standing of blood circulation in veins, he embarked on astudy of the flow of water in narrow capillary tubes undercarefully controlled conditions. Using a sophisticatedlaboratory setup, Poiseuille studied the flow of water inhorizontal capillary tubes varying in diameter from 50m to 150 m and measured fluxes as low as 0.1 cm3

    over several hours. In the absence of gravity, he foundthat water flux was directly proportional to the pressuredifference between the inlet and the outlet and inversely

    proportional to the length of the capillary. These obser-vations were very similar to those made by Ohm in thecase of galvanic current. Although the work was com-pleted in 1842, it was not published until a few yearslater [Poiseuille, 1846]. Similar observations had beenmade earlier in Germany by Hagen [1839].

    One of the most influential works on the flow of waterin porous media during the nineteenth century was thatofDarcy[1856]. Henry Darcy (18031858) was a highlyrecognized civil engineer who is credited with designingand completing, in 1840, the first ever protected townwater supply system in the world. Dissatisfied with theunhealthy sources of drinking water in his native town ofDijon, he helped bring and distribute water from aperennial spring located several kilometers away fromthe town. Later, presumably to build a water purificationsystem, Darcy conducted a series of experiments in ver-tical sand columns to develop a quantitative relationshipfor estimating the rate of flow of water through sandfilters. Darcys experiment was novel in that it includedgravity and involved a natural material (sand) ratherthan an engineered material such as a capillary tube. Hetoo, like Ohm and Poiseuille before him, found that theflux through the column was directly proportional to thedrop in potentiometric head, h z , where z is

    elevation with reference to datum and is pressurehead, directly proportional to the area of cross sectionand inversely proportional to the length of the column.Darcys law plays a fundamental role in many branchesof Earth sciences such as hydrogeology, geophysics, pe-troleum engineering, soil science, and geotechnical en-gineering.

    During the middle of the nineteenth century, poten-

    tial theory was recognized as providing a useful concep-tual-mathematical basis for understanding artesian wellsand other manifestations of deep groundwater circula-tion in France and elsewhere in Europe.

    Soon the heat conduction model of Fourier began tobe used for analyzing circulation of water in groundwa-ter basins. The earliest studies in this regard restrictedthemselves to the steady motion of groundwater. Unlikethe problems of electricity and molecular diffusion, theproblems of groundwater involved large spatial scales(many tens or even hundreds of kilometers laterally andhundreds of meters vertically). Two of the most distin-guished engineers of this era were Jules-Juvenal Dupuit

    (18041866) in France and Philipp Forchheimer (18521933) in Austria. Dupuit [1863] developed the basictheoretical framework for analysis of flow in groundwa-ter systems and of the flow of water to wells.Forchheimer[1886] formally stated the steady seepage of water interms of the Laplace equation and initiated the use ofcomplex variable theory to the solution of two-dimen-sional problems in flow domains of complicated geom-etry that occur in the vicinity of dams and other engi-neering structures. He also pioneered the use of flownets as practical, graphical means of solving seepageproblems in complex flow domains.

    5.3.2. Flow of multiple fluid phases. A significantdevelopment in the study of flow in porous media wasthe work of Edgar Buckingham (18671940). From 1902to 1905 he was an Assistant Physicist with the Bureau ofSoils, U.S. Department of Agriculture. In this brief pe-riod he not only introduced himself to a totally new field,the science of soils, but also made one of the mostimportant contributions to soil physics in particular andthe study of multiphase fluid flow in general. Soon after,he moved to the newly formed National Bureau ofStandards (now National Institute of Standards andTechnology) and became well known for, among otherachievements, his theorem of dimensional analysis.

    In soils close to the land surface, where plant rootsthrive, both water and air coexist. An understanding ofthe dynamics of the occurrence and movement of mois-ture in the soil is critical to judicious agricultural man-agement. When water and air coexist in soils, the con-tacts between air and water in the minute pores arecurved menisci in which energy is stored. As a result, thepressure in the water phase is less than that in the airphase and the difference is the capillary pressure. Thephysics and the mathematics of capillarity had beenenunciated a hundred years earlier by Laplace and byThomas Young (17731829). Buckingham [1907]

    37, 1 / REVIEWS OF GEOPHYSICS Narasimhan: FOURIERS EQUATION 163

  • 7/26/2019 Fouriers Heat Conduction Equation

    14/22

    brought together the work on capillary pressure withthat of Fourier and Ohm on diffusion, and defined thecapillary potential in the water phase as a sum of work tobe done per unit mass against gravity and fluid pressure.He stated that moisture moves in soils in response tospatial variations in potential and that moisture fluxdensity is directly proportional to the gradient of capil-lary potential, the proportionality constant being hy-

    draulic conductivity. Although this statement resemblesthe laws of Fourier, Ohm, and Darcy, there exists a veryimportant difference. In soils that contain water and air,the capillary potential is directly related to water satu-ration, and as water saturation decreases, the flow pathsavailable for moisture movement decrease. Thereforethe conductivity parameter is strongly dependent oncapillary potential, instead of being constant or nearly soas is the case with the laws of Fourier, Ohm, or Darcy.The strong dependence of hydraulic conductivity oncapillary potential renders the study of moisture diffu-sion in soils a very difficult mathematical problem. In anearlier work, Buckingham [1904] also applied the diffu-

    sion equation to the migration of gas in soils and ana-lyzed the dynamic vertical migration of air from the landsurface to the water table in response to fluctuatingatmospheric pressure. Buckinghams work helped re-solve a contemporary paradox in agriculture. In aridregions, where evaporation rates are very high, the soilsare found to be wetter and hold their moisture for muchlonger periods than do the soils of humid areas in dryseasons. Part of the reason for this counterintuitiveobservation is to be found in the dependence of hydrau-lic conductivity on capillary potential, or, equivalently,water saturation. In arid areas, as evaporation rapidly

    desaturates the uppermost soil, the hydraulic conductiv-ity drops practically to zero and further evaporative lossfrom deeper zones is virtually eliminated.

    Note that specific heat, originally defined and mea-sured by Lavoisier and Laplace [1783], is an extremelyimportant physical attribute of materials and occupiesan important position in the transient heat conductionequation. It influences the rapidity with which a materialwill respond thermally to externally imposed perturba-tions: the smaller the capacitance, the faster the re-sponse. Analogously, in the phenomenon of fluid flow inporous media, hydraulic capacitance plays a very impor-tant role. Indeed, the slope of the variation of watersaturation as a function of capillary potential contributesto the hydraulic capacitance of a soil. As a consequence,Buckinghams work lies at the foundation of the dynam-ics of multiple fluid phases in porous media.

    Although Buckingham defined a capillary potentialtheoretically, he could measure it only indirectly in ver-tical columns in which water moves down solely bygravity. He recognized that new instruments would haveto be developed to measure capillary potential underdynamic conditions of flow. Such an instrument wasinvented over a decade later by Willard Gardner (18831964). This ingenious instrument is called the tensiom-

    eter. The key component of this device is a porousceramic cup that is completely saturated with water.Such a porous cup acts like a semipermeable membrane,allowing the flow of water from the soil into the cup, butnot allowing the flow of air. The cup is connected to along, water-filled tube, which is connected to a manom-eter. The tensiometer is set into a natural soil, andthrough exchange of water between the soil and the cup,

    fluid pressure inside of the cup is allowed to attainequilibrium with that in the soil. The equilibrium pres-sure represents the capillary potential. The first mea-surements from this instrument were reported byGard-

    ner et al. [1922]. This strong dependence capacitance oncapillary potential introduces a strong nonlinearity intothe differential equation of moisture movement in soils.

    5.3.3. Hydraulic capacitance. The attribute ofhydraulic capacitance of a naturally occurring porousmaterial such as a soil or a rock arises also for reasonsother than the rate of change of saturation with poten-tial. Earth materials are deformable in response tochanges in the stresses which act on the porous skeleton.

    The ensuing rate of change of pore volume (which isoccupied by water) in response to changes in fluid po-tential also contributes to hydraulic capacitance. Themeasurement of pore volume as a function of fluidpotential was elucidated through the work of Karl Ter-zaghi (18831963), who founded the discipline of soilmechanics. In presenting his experimental results on thedeformation of water saturated clays, Terzaghi [1925]postulated that in water-saturated earth materials,change in pore volume is to be related to the differencebetween skeletal stresses and water pressure. Thus whenskeletal stresses remain unchanged, volume change is

    directly attributable to change in fluid pressure or,equivalently, the fluid potential. Extensive experimentalwork following Terzaghi has shown that earth materialsinvariably exhibit nonelastic deformation behavior.

    In addition, a third component also contributes tohydraulic capacitance of porous materials, the compress-ibility of the fluid itself. Thus hydraulic capacitance innatural geological materials arises from the ability of theporous medium to deform as a result of changing fluidpressure, the ability of the fluid to dilate, and the de-saturation of the pores due to changing capillary pres-sure. Whereas the specific heat of most known materialsdoes not vary by more than a factor of a hundred it is notuncommon for the hydraulic capacitance of soils androcks to vary by a factor of a million.

    If we now look at Fouriers diffusion equation as thebasis for the flow of water in soils and rocks, we see atonce that hydraulic conductivity may vary significantly asa function of fluid potential, as does the hydraulic ca-pacitance. Thus, unlike the heat problem, which is gen-erally characterized by a linear differential equation, thediffusion equation pertaining to flow of water in soilsand rocks is characterized by a highly nonlinear differ-ential equation. Drawing upon the contribution of Buck-ingham and Gardner, Lorenzo Richards (19041993)

    164 Narasimhan: FOURIERS EQUATION 37, 1 / REVIEWS OF GEOPHYSICS

  • 7/26/2019 Fouriers Heat Conduction Equation

    15/22

    presented the transient capillary conduction of water inporous media [Richards, 1931],

    K sA

    t , (4)

    where gz is the potential in the gravity field (inwhich g is gravitational acceleration and z is elevation

    above datum), dp/ is capillary potential (wherep is pressure in the water phase and is density ofwater),

    sis dry bulk density of the soil, andA is the rate

    of change of moisture content with respect to capillarypotential, referred to as the capillary capacity of themedium.

    5.3.4. Hydrodynamic dispersion. Geoffrey Tay-lor (18861975) who studied the advective transfer(transport by the bulk movement of water) of dissolvedsolutes by water in thin capillary tubes made an inter-esting conceptual-mathematical addition to Fouriersdiffusion equation. In a tube through which water isflowing, the velocity of water is practically zero at the

    walls and is at a maximum along the central axis of thetube. Thus although we may be satisfied with an averagevelocity to quantify the total flux of water in the tube, wecannot ignore the velocity variation within the tube if wewish to understand the migration of a solute dissolved inwater. The process is complicated by molecular diffusionwhich will cause the solute to spread perpendicular tothe direction of advective transport. By an elegant math-ematical analysis of the problem, Taylor [1953] showedthat after a sufficient period of time the distribution ofthe solute will exhibit a diffusion-like profile along thedirection of flow and that the effective diffusion coeffi-

    cient is a function of the average velocity as well as thegeometrical attributes of the capillary tube. Recall that

    Fick [1855a, b] considered, in a similar fashion, thevariation of concentration as a function of capillaryradius in osmotic membranes.

    Taylors work inspired the concept of hydrodynamicdispersion, widely used to analyze the migration of con-taminants in groundwater systems. Hydrodynamic dis-persion is a macroscopic diffusion-like process by whichcontaminants dissolved in water in a porous mediumspread (principally along the general flow direction)owing to random variations in flow velocity on a micro-scopic scale.

    In considering the migration of solutes with movingwater in porous materials, it is important to focus onanother attribute of these systems that has direct rele-vance to the heat diffusion equation. Many solutes thatoccur in groundwater also have affinities for the solidsurface. Hence they tend to partition themselves be-tween the aqueous phase and the solid surface by aprocess of adsorption. Adsorption, in turn, is propor-tional to concentration in the aqueous phase. When onewrites the molecular diffusion equation only for theaqueous phase in such systems, the portion of the solutetaken up by the solid surface is accommodated in the

    form of a chemical capacitance term, usually referred toas the retardation factor because this process effectivelyslows down spreading caused by the advective process.In the mathematics of diffusion, the retardation factorplays the same role as specific heat in the heat diffusionequation.

    5.4. Stochastic Diffusion

    5.4.1. Random walk. It is evident from the fore-going discussion that the analyses of the flow of electriccurrent; diffusion in liquids, solids, and gases; and theflow of fluids in porous materials were all directly influ-enced by Fouriers heat conduction model during thenineteenth century. In these analyses, Fouriers modelwas used in an empirical way, to interpret experimentaldata from macroscopic systems. In marked contrast tothis empirical transfer of concept, the second half of thenineteenth century saw the extension of the heat diffu-sion equation to problems of a more theoretical nature,involving the gross manifestation of random processes.

    During the early twentieth century this extension wouldlead to the development of a new field of considerableinterest, namely, stochastic processes. The beginnings ofstochastic diffusion were latent in the work of four quitedifferent workers: the theory of sound of Lord Rayleigh,the law of error of Edgeworth, the theory of speculationof Bachelier, and the theory of Brownian motion ofEinstein. The appellation random walk to denote theseprocesses was coined by Karl Pearson (18571936), abiometrician. The transition from random walk to sto-chastic differential equations would be catalyzed by thework of Langevin. It is now of interest to consider in

    some detail, the conceptual-mathematical developmentsthat led to the theoretical view of the heat conductionequation as an ensemble manifestation of fine-scale ran-dom processes.

    Pearson[1905, p. 242] sought the help of the journalNaturewith the request, Can any one of your readersrefer me to a work wherein I should find a solution of thefollowing problem, or failing the knowledge of any ex-isting solution provide me with an original one? I shouldbe extremely grateful for aid in this matter. A man startsfrom a point O and walks L yards in a straight line; hethen turns through any angle whatever and walks an-other L yards in a second straight line. He repeats thisprocessn times. I require the probability that after these

    nstretches he is