IPhO 2026 · Theory problem 3

Chasing the absolute zero

QUESTION T3 [10 pts]. Chasing the absolute zero

A refrigeration cycle must be carried out by a system with characteristics appropriate for the temperatures to be achieved. One hundred years ago, in 1926, the chemist William Giauque and, independently, the physicist Peter Debye proposed paramagnetic salts as ideal systems for lowering temperatures far below 1 K. Indeed, beginning in the following decade, temperatures as low as 0.0044 K were achieved (Wander de Haas and Gerrit Wiersma, 1935, Leiden Cryogenic Laboratories). In this question, the paramagnetic Carnot refrigeration cycle is studied.

Figure 1

Fig. 3a.

Consider a torus with mean radius \(R\) and inner radius \(r\) made of a homogeneous, isotropic paramagnetic material, around which an insulated conducting wire is wound, with its ends connected to an external voltage source (emf), as shown in figure 3a. The resistance of the wire is so low that energy losses due to the heating of the wire can be neglected. The winding is dense, with \(N\) turns in total, and it is assumed that \(r\ll R\), so that the fields \(\vec H\) and \(\vec B\) and the magnetization \(\vec M\) have approximately constant magnitudes throughout the torus. Let \(V\) and \(A\) be the volume and the cross-section area of the torus. Hereinafter, the abbreviation Pm-T is used to refer to the paramagnetic torus.

In paramagnetic materials, \(\vec M\) is parallel to \(\vec H\) and the following relation holds:

\[ \vec B=\mu_0\vec H+\mu_0\vec M. \]

Energy transfer sign convention: in this problem, Work (\(W\)) and Heat (\(Q\)) are considered positive when they flow into the system, and negative when they flow out.

T3-A [1.0 pts]. Work in the Pm-T

Hint: in magnetic materials, \(\displaystyle\oint_C\vec H\cdot d\vec\ell=I_C\), where \(I_C\) is the net free current passing through the area bounded by the closed curve \(C\).

T3-A1 [0.2 pts]. Let \(H\) be the magnitude of \(\vec H\) in the torus. Write \(H\) in terms of \(N\), \(A\), \(V\) and the instantaneous electric current \(I\) in the wire.

T3-A2 [0.6 pts]. Let \(dW_{emf}\) be the work that the external voltage source performs to change the magnitude of \(\vec B\) by \(dB\). Write \(dW_{emf}\) in terms of \(V\), \(H\), and \(dB\).

T3-A3 [0.2 pts]. The total work done by the voltage source, \(dW_{emf}\), can be divided in two parts: the work that would be performed to change the magnetic field if the toroid had a vacuum core, \(dW_{vac}\), and the work done on the paramagnetic material itself, \(dW\). Write \(dW\) in terms of \(\mu_0\), \(H\), \(V\) and \(dM\).

T3-B [3.0 pts]. Heat and temperature in the Pm-T

The following table compares two thermodynamic systems: the ideal gas and the Pm-T. Use this information and your answer to question T3-A3 to solve the following questions. This information will also be used in part C. If you did not find an answer for \(dW\) in question T3-A3, you may use \(dW=\alpha V\mu_0H\,dM\) where \(\alpha\) is a number. The volume \(V\) of the paramagnetic torus is constant, so that there is no work due to pressure. Below, \(T\) is the temperature in all instances.

IDEAL GAS PARAMAGNETIC TORUS
State functions: \(P,V\) and temperature \(T\) State functions: \(H,M\) and temperature \(T\)
Constants: \(R,\gamma\) and number of moles \(n\) Constants: \(K,\lambda\) and number of moles \(n\)
Equation of state: \(PV=nRT\) Equation of state: \(TMV=nKH\)
Heat capacity at constant \(V\): \(C_V=nR/(\gamma-1)\) Heat capacity at constant \(M\): \(C_M=n\lambda/T^2\)
Internal energy \(U\) in any process satisfies \(dU=C_V\,dT\) Internal energy \(U\) in any process satisfies \(dU=C_M\,dT\)

\(K\) and \(\lambda\) are constants given by the material in the torus. You may use the equations provided in the table directly without proving them.

T3-B1 [1.5 pts]. If the magnitude of \(\vec H\) changes from \(H_i\) to \(H_f\) at a constant temperature \(T\), an energy \(Q\) is transferred between the Pm-T and its environment. Write \(Q\) in terms of \(\mu_0\), \(n\), \(V\), \(K\), \(H_i\), \(H_f\), and \(T\).

T3-B2 [1.5 pts]. If the magnitude of \(\vec H\) changes adiabatically from \(H_i\) to \(H_f\), the temperature of the Pm-T changes from \(T_i\) to \(T_f\). Write \(\Delta T=T_f-T_i\) in terms of \(\mu_0\), \(V\), \(n\), \(K\), \(\lambda\), \(H_i\), \(H_f\), and \(T_i\).

T3-C [6.0 pts]. Carnot refrigerator of the Pm-T

Figure 2

Fig.3b.

The \(H\) vs. \(T\) diagram of the Carnot refrigeration cycle \(1\to2\to3\to4\to1\) performed by the Pm-T is shown in figure 3b. In this cycle, let:

  • \(T_h\): temperature of the hot reservoir
  • \(T_c\): temperature of the cold reservoir
  • \(Q_h\): absolute value of the heat transferred to the hot reservoir
  • \(Q_c\): absolute value of the heat absorbed from the cold reservoir

T3-C1 [0.2 pts]. Mark \(T_h\) and \(T_c\) on the \(T\) axis of the \(H\) vs. \(T\) diagram. Also label the processes in which the transfers of \(Q_h\) and \(Q_c\) occur, respectively.

T3-C2 [1.5 pts]. Let \(M_1\), \(M_2\), \(M_3\), and \(M_4\) be the magnitudes of \(\vec M\) at the vertices of this cycle. Write \(M_1\) in terms of \(M_2\), \(M_3\), and \(M_4\).

T3-C3 [0.8 pts]. Suppose that to cool 1.00 L of liquid helium initially at 1.00 K, we utilize a 2.0 moles Pm-T made of potassium chromate. For this material, \(K=1.87\times10^{-6}\ \mathrm{K\cdot m^3/mol}\), its density is \(2730\ \mathrm{kg/m^3}\) and its molar mass is \(0.19\ \mathrm{kg/mole}\). The Pm-T performs a Carnot refrigeration cycle with \(H_1=411624\ \mathrm{A/m}\), \(H_2=311306\ \mathrm{A/m}\), \(H_3=204618\ \mathrm{A/m}\) and \(H_4=240446\ \mathrm{A/m}\). After one operating cycle, calculate the final temperature of the liquid helium. Assume that for the liquid helium both the specific heat capacity \(c=100\ \mathrm{J/(kg\cdot K)}\) and the density \(\rho=130\ \mathrm{kg/m^3}\) remain constant.

T3-C4 [2.0 pts]. In order to continuously lower the temperature of a body with constant heat capacity \(C_c\), Carnot cycles are allowed to occur over time in such a way that the power \(P\) transferred to the refrigerator remains constant. It is then assumed that the temperature \(T_h\) of the hot reservoir remains constant and that in each operating cycle the temperature \(T_c\) of the cooled body decreases by \(dT_c\), satisfying the relation \(dQ_c/dQ_h=T_c/T_h\). If at \(t=0\) we have \(T_c=T_0\), determine the time \(t\) that the Carnot refrigerator must operate for the cooled body to reach a temperature \(T<T_0\). Express your answer in terms of \(C_c\), \(T_h\), \(P\), \(T_0\) and \(T\).

T3-C5 [1.5 pts]. The efficiency of a refrigerator is quantified by its coefficient of performance \(\mathrm{COP}=Q_c/W\). Determine the overall COP of the refrigerator of question T3-C4 with all the cycles performed until time \(t\). Write your answer in terms of \(T_0\) and \(T_h\) and \(T\).

Physical Constants

Reference values for selected physical constants.

Constant Notation Value Units
Speed of light in vacuum \(c\) \(2.99792458\times10^8\) \(\mathrm{m\cdot s^{-1}}\)
Planck constant \(h\) \(4.135667696\ldots\times10^{-15}\) \(\mathrm{eV\cdot s}\)
Reduced Planck constant \(\hbar=h/2\pi\) \(6.582119569\ldots\times10^{-16}\) \(\mathrm{eV\cdot s}\)
Avogadro number \(N_A\) \(6.02214076\times10^{23}\) \(\mathrm{mol^{-1}}\)
Boltzmann constant \(k_B\) \(8.617333262\times10^{-5}\) \(\mathrm{eV\cdot K^{-1}}\)
Ideal gas constant \(R\) \(8.31446261815324\) \(\mathrm{kg\cdot m^2\cdot s^{-2}\cdot mol^{-1}\cdot K^{-1}}\)
Elementary charge \(e\) \(1.602176634\times10^{-19}\) \(\mathrm{A\cdot s}\)
Universal gravitational constant \(G\) \(6.67430(15)\times10^{-11}\) \(\mathrm{m^3\cdot kg^{-1}\cdot s^{-2}}\)
Stefan Boltzmann constant \(\sigma\) \(5.670374419\times10^{-8}\) \(\mathrm{kg\cdot s^{-3}\cdot K^{-4}}\)
Vacuum permeability \(\mu_0\) \(1.25663706127(20)\times10^{-6}\) \(\mathrm{kg\cdot m\cdot A^{-2}\cdot s^{-2}}\)
Vacuum permittivity \(\varepsilon_0\) \(8.8541878188(14)\times10^{-12}\) \(\mathrm{A^2\cdot s^4\cdot kg^{-1}\cdot m^{-3}}\)
Mass of the electron \(m_e\) \(5.1099895069(16)\times10^5\) \(\mathrm{eV}/c^2\)
Mass of the proton \(m_p\) \(9.3827208943(29)\times10^8\) \(\mathrm{eV}/c^2\)
Atomic mass units \(\mathrm{amu}\) \(9.3149410372(29)\times10^8\) \(\mathrm{eV}/c^2\)

Official English text of the problem.

Answers

  • T3-A1: \[ \boxed{H=\frac{NAI}{V}}. \]

  • T3-A2: \[ \boxed{dW_{emf}=VH\,dB}. \]

  • T3-A3: \[ \boxed{dW=\mu_0VH\,dM},\qquad \alpha=1. \]

  • T3-B1: \[ \boxed{Q=\frac{\mu_0nK}{2T}\left(H_i^2-H_f^2\right)}. \]

  • T3-B2: \[ \boxed{\Delta T =T_i\left[ \sqrt{\frac{\lambda+\mu_0KH_f^2} {\lambda+\mu_0KH_i^2}}-1 \right]}. \]

  • T3-C1: The left vertical line, through states \(2,3\), is at \(T_c\); the right vertical line, through states \(1,4\), is at \(T_h\). Heat \(Q_c\) is absorbed during \(2\to3\), and heat \(Q_h\) is rejected during \(4\to1\).

  • T3-C2: \[ \boxed{M_1=\sqrt{M_2^2+M_4^2-M_3^2}}. \]

  • T3-C3: Using \(T_c\simeq1.00\ \mathrm K\) for the cold stroke, \[ \boxed{T_{\mathrm{He},f}\simeq0.990\ \mathrm K}. \] This is a small-cooling estimate that takes the cold isotherm at the initial helium temperature; it uses only the cold stroke \(2\to3\) (see the note at the end of the T3-C3 solution).

  • T3-C4: \[ \boxed{t=\frac{C_c}{P} \left[T_h\ln\!\left(\frac{T_0}{T}\right)-(T_0-T)\right]}. \]

  • T3-C5: \[ \boxed{\mathrm{COP}_{\mathrm{overall}} =\frac{T_0-T} {T_h\ln(T_0/T)-(T_0-T)}}. \]

Solution

We use the prescribed sign convention throughout: \[ dU=dQ+dW, \] where heat and work are positive when supplied to the paramagnetic material. The magnetic processes are treated quasistatically, with the uniform-field approximation \(r\ll R\) and negligible resistive losses.

T3-A1

Choose an Amperian circle of radius \(R\) inside the torus. It links all \(N\) turns, so the enclosed free current is \(NI\). Thus, \[ H(2\pi R)=NI. \] Since the torus volume is \(V=(2\pi R)A\), \[ \boxed{H=\frac{NI}{2\pi R}=\frac{NAI}{V}}. \]

T3-A2

The magnetic flux through each turn is approximately \[ \Phi=AB. \] Faraday’s law gives an induced emf \(-NA\,dB/dt\). With negligible resistance, the applied source emf balances this: \[ \mathcal E=NA\frac{dB}{dt}. \] The electrical work supplied in time \(dt\) is therefore \[ dW_{emf}=\mathcal E I\,dt=NAI\,dB. \] Using \(NAI=VH\) from A1, \[ \boxed{dW_{emf}=VH\,dB}. \]

This expression also gives the correct sign when the source receives energy during a decrease of the field.

T3-A3

Because the fields and magnetization are parallel, \[ dB=\mu_0\,dH+\mu_0\,dM. \] Consequently, \[ dW_{emf} =\mu_0VH\,dH+\mu_0VH\,dM. \]

For a vacuum core, \(B=\mu_0H\), so the work associated with the same change of \(H\) is \[ dW_{vac}=\mu_0VH\,dH =d\!\left(\frac{\mu_0VH^2}{2}\right). \] Subtracting this vacuum-field contribution leaves the work on the material: \[ \boxed{dW=dW_{emf}-dW_{vac}=\mu_0VH\,dM}. \] Thus the fallback coefficient is \(\alpha=1\).

T3-B1

The given internal-energy relation implies that an isothermal process has \[ dU=\frac{n\lambda}{T^2}\,dT=0. \] The first law then gives \[ dQ=-dW=-\mu_0VH\,dM. \]

At constant \(T\), the equation of state yields \[ M=\frac{nKH}{VT}, \qquad dM=\frac{nK}{VT}\,dH. \] Hence \[ Q=-\frac{\mu_0nK}{T}\int_{H_i}^{H_f}H\,dH, \] and therefore \[ \boxed{Q=\frac{\mu_0nK}{2T}(H_i^2-H_f^2)}. \]

The volume cancels. Isothermal demagnetization, \(H_f<H_i\), absorbs heat; isothermal magnetization releases heat.

T3-B2

An adiabatic process has \(dQ=0\), so \[ \frac{n\lambda}{T^2}\,dT=\mu_0VH\,dM. \] Now both \(H\) and \(T\) vary. Differentiating the equation of state gives \[ dM=\frac{nK}{V} \left(\frac{dH}{T}-\frac{H}{T^2}\,dT\right). \] Substitution into the first law gives \[ \frac{n\lambda}{T^2}\,dT =\mu_0nK \left(\frac{H}{T}\,dH-\frac{H^2}{T^2}\,dT\right). \] Thus \[ (\lambda+\mu_0KH^2)\,dT =\mu_0KHT\,dH, \] or \[ \frac{dT}{T} =\frac{\mu_0KH}{\lambda+\mu_0KH^2}\,dH. \]

Integrating between the initial and final states, \[ \ln\frac{T_f}{T_i} =\frac12\ln \frac{\lambda+\mu_0KH_f^2} {\lambda+\mu_0KH_i^2}. \] Therefore, \[ T_f=T_i \sqrt{\frac{\lambda+\mu_0KH_f^2} {\lambda+\mu_0KH_i^2}}, \] and \[ \boxed{\Delta T =T_i\left[ \sqrt{\frac{\lambda+\mu_0KH_f^2} {\lambda+\mu_0KH_i^2}}-1 \right]}. \]

For positive \(K\) and heat capacity, decreasing \(H\) decreases \(T\): this is adiabatic demagnetization.

An equivalent form, useful below, is the adiabatic invariant \[ \boxed{\frac{\lambda+\mu_0KH^2}{T^2} =\text{constant along an adiabat}}. \]

T3-C1

The horizontal coordinate is temperature. Therefore:

  • Label the left dashed projection, beneath states \(2\) and \(3\), \(T_c\).
  • Label the right dashed projection, beneath states \(1\) and \(4\), \(T_h\).

On \(2\to3\), the field decreases at constant \(T_c\). By B1, the torus absorbs heat, so place \(Q_c\), directed into the torus, beside this downward vertical stroke.

On \(4\to1\), the field increases at constant \(T_h\). The torus releases heat, so place \(Q_h\), directed out of the torus, beside this upward vertical stroke.

The remaining strokes are \(1\to2\), adiabatic cooling, and \(3\to4\), adiabatic heating.

T3-C2

Substitute \[ H=\frac{VTM}{nK} \] into the adiabatic invariant from B2: \[ \frac{\lambda}{T^2} +\frac{\mu_0V^2}{n^2K}M^2 =\text{constant}. \]

For the adiabat \(1\to2\), \[ \frac{\mu_0V^2}{n^2K}(M_1^2-M_2^2) =\lambda\left(\frac1{T_c^2}-\frac1{T_h^2}\right). \] For the adiabat \(3\to4\), \[ \frac{\mu_0V^2}{n^2K}(M_4^2-M_3^2) =\lambda\left(\frac1{T_c^2}-\frac1{T_h^2}\right). \] Their right-hand sides are identical, so \[ M_1^2-M_2^2=M_4^2-M_3^2. \] Since the \(M_j\) are magnitudes, the positive root is required: \[ \boxed{M_1=\sqrt{M_2^2+M_4^2-M_3^2}}. \]

T3-C3

For the one-cycle estimate, take the cold-stroke temperature as approximately the initial helium temperature: \[ T_c\simeq1.00\ \mathrm K. \] The heat absorbed during \(2\to3\) is \[ Q_c=\frac{\mu_0nK}{2T_c}(H_2^2-H_3^2). \] Using the supplied values, \[ Q_c= \frac{ (1.25663706127\times10^{-6})(2.0)(1.87\times10^{-6}) }{ 2(1.00) } \left(311306^2-204618^2\right)\ \mathrm J, \] which gives \[ Q_c=0.1293459\ldots\ \mathrm J\simeq0.13\ \mathrm J. \]

The helium mass and heat capacity are \[ m_{\mathrm{He}} =\rho V_{\mathrm{He}} =(130)(1.00\times10^{-3}) =0.130\ \mathrm{kg}, \] \[ C_{\mathrm{He}}=m_{\mathrm{He}}c =(0.130)(100)=13.0\ \mathrm{J\,K^{-1}}. \] Energy conservation for the helium gives \[ C_{\mathrm{He}}(T_{\mathrm{He},f}-1.00\ \mathrm K)=-Q_c. \] Thus \[ T_{\mathrm{He},f} =1.00\ \mathrm K-\frac{0.1293459\ldots\ \mathrm J} {13.0\ \mathrm{J\,K^{-1}}} =0.9900503\ldots\ \mathrm K, \] or \[ \boxed{T_{\mathrm{He},f}\simeq0.990\ \mathrm K}. \]

The estimated temperature decrease is approximately \(1\%\), supporting the small-change approximation for this numerical estimate. The material volume cancels from \(Q_c\), so its density and molar mass are unnecessary here. For an exact cycle the four given fields are not fully consistent: with \(M=nKH/(VT)\) on the two isotherms, the relation of T3-C2 requires \(T_h/T_c=\sqrt{(H_1^2-H_4^2)/(H_2^2-H_3^2)}\approx1.42\), and the adiabats \(1\to2\) and \(3\to4\) then hold only with \(\lambda\approx-0.062\ \mathrm{J\,K\,mol^{-1}}\), that is, a negative heat capacity \(C_M=n\lambda/T^2\). The estimate above uses only the cold stroke and does not depend on \(\lambda\).

T3-C4

Let \(dT_c<0\) denote the actual temperature change of the cooled body. The positive heat extracted from it is \[ dQ_c=-C_c\,dT_c. \] The specified Carnot relation gives \[ dQ_h=\frac{T_h}{T_c}\,dQ_c. \] Since the refrigerator returns to its original state after each cycle, its net internal-energy change vanishes. The required work is therefore \[ dW=dQ_h-dQ_c =\left(\frac{T_h}{T_c}-1\right)dQ_c. \] With constant input power, \[ P\,dt=dW =-C_c\left(\frac{T_h}{T_c}-1\right)dT_c. \]

Integrating from \(T_c=T_0\) at \(t=0\) to \(T_c=T\), \[ Pt=-C_c\int_{T_0}^{T} \left(\frac{T_h}{T_c}-1\right)dT_c. \] Hence \[ Pt=C_c \left[ T_h\ln\!\left(\frac{T_0}{T}\right)+T-T_0 \right], \] and \[ \boxed{ t=\frac{C_c}{P} \left[ T_h\ln\!\left(\frac{T_0}{T}\right)-(T_0-T) \right]}. \]

For \(0<T<T_0<T_h\), this time is positive. Moreover, \[ T\to0^+\quad\Longrightarrow\quad t\to\infty. \] Within the stated model, reaching absolute zero requires an infinite operating time.

T3-C5

The total heat extracted from the cooled body is \[ Q_{c,\mathrm{total}}=C_c(T_0-T). \] The total work supplied is \[ W_{\mathrm{total}}=Pt =C_c\left[ T_h\ln\!\left(\frac{T_0}{T}\right)-(T_0-T) \right]. \] Therefore, \[ \boxed{ \mathrm{COP}_{\mathrm{overall}} =\frac{Q_{c,\mathrm{total}}}{W_{\mathrm{total}}} =\frac{T_0-T} {T_h\ln(T_0/T)-(T_0-T)} }. \]

As a check, for an infinitesimal temperature decrease, \[ \lim_{T\to T_0^-}\mathrm{COP}_{\mathrm{overall}} =\frac{T_0}{T_h-T_0}, \] which is the usual Carnot refrigerator COP at cold temperature \(T_0\).

The solution was checked against the official solution.