Classroom

In this section of Resonance, we invite readers to pose questions likely to be raised in a classroom situation. We may suggest strategies for dealing with them, or invite responses, or both. Classroom is equally a forum for raising broader issues and sharing personal experiences and viewpoints on matters related to teaching and learning science. Getting Acquainted with Gears and Wheels Quantum Mechanically
Kamal Sharma and N Kumar Raman Research Institute Bangalore 560 080, India. Email: nkumar@rri.res.in

For a pair of wheels or gears with positive coupling, i.e., without slip or play, there are rules of engagement that have some interesting consequences when their dynamics is treated quantum mechanically. We will illustrate the principal ideas involved here with the help of an elementary, basically a textbook exercise whose solution, however, is not only interesting, but may also be re-interpreted rather creatively. Possible relevance of this simple exercise to the incredible, ever-shrinking world of the nano (1 nm = 109 m) is pointed out. 1. Introduction and Motivation The invention of the wheel has been considered by many to be one of the oldest and the greatest inventions of mankind, dating back to ca.1000 BC. The same may more are less, be said of the gear (Figure 1) { after all, the toothless wheel is an analog of the digital toothed gear. For a mathematically minded number theorist, however, we hasten to add that the gear has a countable discrete-

Keywords Nano-gears, no-slip condition, constraint, quantum mechanics.

RESONANCE January 2013

67

CLASSROOM

Figure 1. Two gears engaged so as to transmit motion.

ness, i.e., its teeth can be counted, whereas the wheel is an uncountable continuum. The twain have dierent cardinality! In the following, however, we will continue to use the two terms interchangeably. After all, for any wheel one can always nd a gear approximant to the desired order of precision/resolution. Some of these issues will recur in the course of our discussion. Over the millennia, natural philosophers (physicists and mathematicians) and engineers alike have engaged with the gear/wheel and the complex mechanisms based on these. In fact, Aristotle (ca. 300 BC) had suggested the use of the frictionally engaged wheels to transmit motion. Some of the other great names that readily come to the mind are Archimedes (ca. 250 BC) for his association with the design of the antikythera { a geared mechanical computer of great complexity; Galileo (ca. 1600 AD) for the Galilean Jovilabe with the surprising non-reversibly geared clock mechanism; and so on until we nally come to James Clerk Maxwell (1837{ 1879), who created a truly powerful mental construct, a mechanical model for the electromagnetic phenomena { with tiny spinning cells, idler wheels, and roller contacts without slip, lling the space [1]. We have, of course, come a long way { from the potter's crude wooden wheels to the nely geared watches and the robust power machines { and we see gears at every turn of an engine. But, no matter how sophisticated and varied are our gears today as compared to the potter's ancient wheels, they all share a basic commonality { they are all governed by classical mechanics (Newton's Laws of Motion). But, this may have to change now as we continually downsize our devices to the nano/subnanoscales; and still there seems to be plenty of room left at the bottom [2]. A question one may reasonably ask now is whether these enmeshed nano-gears will behave essentially in the same manner as the gears we are familiar with. Well, the answer is a denite NO!

68

RESONANCE January 2013

CLASSROOM

Figure 2. Two engaged wheels without slip; analog version of the enmeshed digital gear.

The domain of the nano/sub-nanoscale is ruled by the quantum. There may be surprises and even some creative re-interpretations possible. We must, therefore, get re-acquainted with the gear and the wheel { but now quantum-mechanically! And we will do this here through a study of the simplest of examples, viz., a pair of coupled circular discs lying in a plane. Consider a system of two co-planar discs 1 and 2 of radii a1 and a2, free to rotate about their respective centres c1 and c2 xed in the plane, and enmeshed such that there is no slip at the point of their contact P as shown in Figure 2. This schematic depicts our elementary model system of two wheels/gears properly engaged. A classical treatment of this two-body dynamical system begins with the Lagrangian, L = K U for the system, where K is its total kinetic energy expressed _1 and _2 , and U is in terms of the (angular) velocities the total potential energy as a function of the (angular) coordinates 1 and 2 . Inasmuch as there is no energy term depending on the (angular) coordinates 1 and 2, the total potential energy U = 0. Now, the total kinetic energy (K ) of the coupled system is given by 1 _ 2 1 _2 K = I1 + I2 2 1 2 2 (1)

RESONANCE January 2013

69

CLASSROOM

1 2 with I1(= 1 M1 a2 1 ) and I2 (= 2 M2 a2 ) being the moments 2 of inertia of the two circular discs of mass M1 and M2 respectively. Here the overhead dot denotes time derivative.

Next, we turn to the no-slip constraint. This may be implemented through the condition _1 = a2 _2 : a1 (2)

The classical dynamics is, therefore, completely determined by the total kinetic energy K and the no-slip constraint, as in (1) and (2) respectively. (A classical dynamical system with constraints involves, in general, somewhat advanced methods of classical mechanics [3]. Further, the presence of any angle-dependent potential would have made the no-slip condition non-trivial to implement). In the present case, however, we can simply rewrite the constraint in equation (2) as, _2 = a1 _1 ; (3) a2 _2 , say, in favour of _1 throughout. and thus eliminate _1 in favour of (Clearly, we could equally well eliminate _2, and then repeat the same treatment as given below. We will obtain the same nal results. The reader is encouraged to verify this explicitly). We can now express _1 , the kinetic energy K given in (1) as a function of alone: " 2 # 1 a1 1 _2 2 _1 K= I1 + I2 I (4) 2 a2 2 1 with I an eective moment of inertia. Thus, the no-slip constraint in (2) has eectively reduced the number of degrees of freedom by one (without changing the physics, of course). For the sake of clarity, let us rename this remaining degree of freedom as ( 1 ). Since, as noted above, there is no potential energy term in this problem, the kinetic energy K expressed entirely in terms

70

RESONANCE January 2013

CLASSROOM

_ 1 ( _ ) is nothing but the total of the angular velocity Lagrangian L for the system. Next, we go from the Lagrangian (L) to the Hamiltonian (H ) which is convenient for treating the problem quantum mechanically. For this, we have to introduce the canonical (angular) momentum J conjugate to the (angular) coordinate as J= @L _ = I : _ @ (5)

The classical Hamiltonian H is then given by the Legen_ L, and expressing the latdre transformation H = J ter in terms of the canonical momentum J using (5). We at once obtain [3] H= J2 : 2I (6)

This essentially completes the classical Hamiltonian formulation of the 2-disc problem with the no-slip constraint built in. It is to be noted that equation (6), together with (5), describes eectively a single -disc problem, with the eective disc moment of inertia = I . This is classically exact in and of itself. Here J is the constant of motion related to the conserved energy (= J 2 =2I ) of the system. We would like to re-emphasize here that in a classical mechanical treatment, the two-disc problem with no-slip condition is mathematically completely equivalent to a one-disc problem; i.e., from their classical dynamical (Newtonian) equations of motion, one cannot tell the two apart. As will be shown below, this, however, is not the case quantum-mechanically, and the physics becomes rather subtle. Let us, therefore, turn now to a quantum treatment of the same two-disc problem. Note rst that in the classical case as treated above, the existence of the other disc (disc 2) is, of course, implicit { but it enters only parametrically in the eective one-disc Hamiltonian as

RESONANCE January 2013

71

CLASSROOM

a modied moment of inertia (equation (6)). In the quantum-mechanical treatment too, the Hamiltonian formally appears as in the one-disc problem, but the mere existence of the (now hidden ) other disc in the original two-disc problem brings about certain global changes in the allowed solutions that have observable consequences quantitatively as well as qualitatively. Let us see how this comes about. In the quantum case, the classical Hamiltonian H and the angular momentum J as derived above are to be ^ and J ^), with treated as operators (H ^2 ^ =J ; H 2I and ^ = i ~ @ ; J @ (7)

whose eigenvalues give the simultaneously allowed energy and the angular momentum for the 2-disc dynami^ and J ^ are Hermitian operators cal system. (Note that H with real eigenvalues, and they commute). All we have to do now is to solve the associated eigenvalue equation (the time-independent Schr odinger equation) ~2 @ 2 () ^ H () = E(); 2I @2 (8)

where we have explicitly indicated the (angular) coordinate as the argument of the wavefunction (). (Recall that 1 ). The solution is readily seen to be () / eik with Ek = ~2 k 2 2I (the energy eigenvalue); (9b) (to within a normalization) (9a)

where the parameter k is to be determined from the boundary condition for the problem as follows. Formally, (9) may well describe a one-disc problem with an eective moment of inertia I1 + I2 (a1=a2 )2 . For
72 RESONANCE January 2013

CLASSROOM

such a one-disc problem, the single-valuedness of the wavefunction would demand a periodic boundary condition with period 2, i.e., () = ( + 2) inasmuch as for the single disc, the angles and + 2 refer to identical single-disc congurations (i.e., a complete cycle, or a 2 -turn restores the disc to status quo ante !). The periodic boundary condition eik2 = 1 then forces k to be an integer (= N , say) giving the allowed en2 2 ergy eigenvalues as EN = ~ 2N . This result is, however, I not valid in the quantum two-disc case inasmuch as the two enmeshed sub-systems, namely the two discs, must be treated as a single whole system with a global wavefunction. The uniqueness of this global wavefunction, therefore, demands that any two congurations of the two-disc system be now identied as being the same if they dier by integral numbers (n1 and n2 , say) of complete turns of the two component discs 1 and 2. This leads to a novel strictly quantum mechanical feature, as discussed below. The no-slip condition now constrains j n1a1 j=j n2a2 j with a1 a2, say, without loss of generality. Clearly, this can be satised if and only if the wheel ratio a1=a2 = a rational number. (This is quite analogous to the problem of periodic occurrences (syzygy ) of straight-line alignment of planets in our solar system, the period being essentially the lowest common multiplier (LCM) of the planetary periods). Thus, in our case, for example, if a1 =a2 = 4=6 2=3, then 3 turns of disc 1 will correspond to 2 turns of disc 2. So, for this case, the single-valuedness of the global wavefunction requires 2 N2 k 3 2 = N 2 , or k = N=3, giving EN = ~ 2I 9 as the allowed energy eigenvalues, where the energy-level index N is an integer. For the general case of rational wheel ratio, therefore, the energy eigenvalues EN turn out to be given by EN = ~2 N 2 ; 2In2 1 (10)

RESONANCE January 2013

73

CLASSROOM

where the wheel ratio a1=a2 = n2 =n1 , and the integers n2 and n1 are without common factors (relatively prime). For gears, the gear ratio is clearly a rational number and the above treatment goes right through. But, on the other hand, for the case of two discs the wheel ratio (analogue of the gear ratio) can be irrational. For this case, we cannot nd integers n1; n2 such that j n1a2 j=j n2 a2 j, as required by the no-slip (constraint) condition. Then the above scheme of quantization clearly fails. It is, however, still possible to construct physically sensible solutions for the rational wheel-ratio approximants to a desired order p of precision. Thus, e.g., for the wheel ratio a1 =a2 = 1= 2, with p 2 = 1:41421 35623 73095 04880 16887 24209 69807 ; clearly an irrational p number, we can construct rational approximants for 2 to increasing orders of precision, namely 14 ; 141 ; 1414 . (Note that these rational ap10 100 1000 proximants have to be listed out in long hand there is no shortcut to the approximant of a given order). In Figure 3, we have plotted semi-logarithmically the dimensionless energy-level values eN (2I=~2 )EN vs the

Figure 3. Energy spectrum for a two-disc system having irrational wheel ratio (ration of radii). Semi-log plot of dimensionless level energy

EN EN (2I=~2 ) vs the level index N for different orders m of the rational approximant.

74

RESONANCE January 2013

CLASSROOM

energy-level index N for various orders (m) of the rational approximants. As can be readily seen, the energy levels tend to a continuum as m tends to innity, which is to be expected physically. It is interesting to note that the above treatment of the no-slip two-disc system can be readily extended to the case of many discs coupled without slip, e.g., a system including a number of idlers. Let us provoke the still interested reader with a tempting generalization in which our two discs are now replaced with two spheres, as shown in Figure 4, free to turn about their xed centres c1 and c2 , but again under the no-slip rule of engagement. Now, however, the no-slip constraint operates entirely dierently { thus, arbitrary rotations of the two spheres about the common z -axis passing through their centres c1 and c2 , subtend no linear velocities in the common tangent plane normal to the c1 c2 axis at the point of contact P, i.e., the no-slip constraint is trivially satised. On the other hand, for rotations about other orthogonal axes, this is not the case and the no-slip condition has to be imposed nontrivially. One has to go to some convenient coordinate system to simplify the calculation. Quantum-mechanically, there is an added complication too while classically, innitesimal rotations commute, quantum mechanically they do not. This is expected to

c1
P

c2

Figure 4. Two engaged spheres free to rotate about their centres without slip.

RESONANCE January 2013

75

CLASSROOM

make the two-sphere case very dierent from its twodisc counterpart when the problem is treated quantum mechanically. The whole point really is that the other (hidden ) disc/ sphere shows up very dierently in the classical and the quantum cases. Thus, in the classical case, the hidden disc/sphere aects the eectively single-disc/sphere problem through a mere modication of the parameters occurring in its energy expression. (This is quite analogous to the case of the invisible Neptune showing up as causing a visible deviation of Uranus from its otherwise expected orbit). In the quantum case, however, the no-slip constraint enters qualitatively dierently through the requirement of single-valuedness of the global wavefunction for the whole system. Thus, the effect of the other, hidden disc/sphere, admits a possible re-interpretation in terms of an eective one-disc/sphere system as being a new elementary object in its own right, e.g., in the case of the two-sphere system, the new object may have a dierent, even fractional, angular momentum! The interested reader is encouraged to explore this rather subtle point further { the unreasonable eectiveness of the hidden variable! Finally, quantum wheels and gears are to be viewed in the present context of the fast emerging world of the nano [6]. Figure 5 demonstrates the idea of a molecularscale train of nano-gears, 1.2 nm across and a few atoms

Figure 5. W orking molecular nano-gear 1.2 nm across and a few atoms wide. From [7].

76

RESONANCE January 2013

CLASSROOM

wide already made, or rather synthesized in the laboratory [7]. The possibility of realizing other machine components, like nano-rods, axles, van der Waals bearings, and the -bond based hinges for the nano, indeed organic-molecular scales, seems quite realistic now. Quantum wheels and gears will make us re-think many other physical aspects of the nano such as friction and lubrication on the nanoscale. The high surfaceto-volume ratio should, for example, help us with the heating problem. We would like to conclude this conversational classroom exercise on the note that it is time now to gear up to the challenges of the nano quantum mechanically!
Suggested Reading
[1] For a thoughtful account of Maxwells creative ideas on a mechanical model, with spinning cells, idler wheels and roller contacts without slip, for the electromagnetic phenomena, see Basil Mahon, The Man Who Changed Everything: The Life of James Clerk Maxwell, John Wiley, 2003. [2] R P Feynman, There is plenty of room at the bottom, Caltech Engg. & Science, Vol.23, pp.2236, 1960. Available at http://www.zyvex.com/nanotech/feynman.html. [3] [4] [5] H Goldstein, Classical Mechanics, Reading, Mass., Addison-Wesley Publishing Co. Inc., 1950. D J Griffiths, Introduction to Quantum Mechanics, V Edition, PrenticeHall of India, New Delhi, 2006. C Kittel, Introduction to Solid State Physics, Wiley Eastern University Press, New Delhi, 1977. Also, see Angus Mackinnon, arXiv: cond-mat/ 020 5647 [6] [7] NANO, The Emerging Science of Nanotechnology, Little, Brown and Company, London, 1995. See ScienceDaily , June 15, 2009. http://www.sciencedaily.com/releases/2009/\\06/090615102036.htm

RESONANCE January 2013

77