YBCO Synthesis and Superconductor Levitation — Solid-State Reaction, Oxygen Content, and Flux Pinning Explained

Back to: Mochizuki Research Group Education — Inorganic Materials Science > Virtual Experiments, Inorganic Materials Science resources

This page is the explanation that pairs with ybco-simulator-en.html. Concerning the solid-state synthesis of YBa2Cu3O7−δ (hereafter YBCO) carried out in basic experiments A1–A3 of the Materials Science Experiments (Inorganic) course, and the observation of magnetic levitation of the sintered body, this page summarizes the reaction equation and mass balance, what happens inside the material during calcination, sintering and oxygenation, the relation between the oxygen content δ and the superconducting transition temperature Tc, and “why it levitates and why its position is stable,” together with the equations used in the simulator and their sources. For the principles of superconductivity itself (GL theory and BCS theory), see superconductivity-simulator-en.html; for Cooper pairs, see cooper-pair-simulator-en.html. How to read the screen of each tab of the simulator is explained in Appendix A.

The backbone of this page The YBCO experiment is easier to follow if you divide it into three stages: ① making the phase (solid-state reaction: calcination), ② making the shape (pressing and sintering: densification), and ③ putting in oxygen (oxygenation: making δ smaller). Whether the sample levitates is not decided only by whether it is “below Tc”; it changes with the superconducting volume, the size over which the shielding current circulates (a grain, or the whole sample), the strength of the flux pinning, and the cooling history.
On checking against the source materials (as of September 25, 2026) The procedures and values (reagent masses; times and temperatures of mixing, drying, calcination and regrinding; die diameter, pressing pressure and holding time; sintering and 400 °C hold conditions; polishing; four-probe current; levitation arrangement and procedure) follow the 2025 edition of the lab manual for Materials Science Experiments (Inorganic Materials) of the former Tokyo Institute of Technology (§5, §6.1–6.6) and the lab videos (A1–A3). Heat treatments are done in air (as specified by the instructor in charge, 2026-09-25). Student experiments are done in air; in research, a tube furnace with flowing N2 or a glove box filled with Ar is sometimes used (as specified by the instructor in charge). Values that appear in neither the manual nor the videos (the cooling after calcination, the duration of dry mixing, the amount of ethanol, the amount of powder used for pressing, the amount removed by polishing, and the material and dimensions of the magnet) are (provisional). Oxygenation in flowing oxygen (450 °C, 12 h) is a condition from the literature and for comparison, not a condition of the lab course. The calculated values that correspond to the data reduction in the manual (theoretical mass loss, theoretical yield, true density and relative density by the manual’s method, and the reading of the hydraulic gauge) are also shown in the text and in the simulator, following Prof. Mochizuki’s reply (2026-09-25). §12 collects the items to be confirmed and how they were resolved.

1. Introduction — The Flow of the Experiment and the Role of the Simulator

The procedure of basic experiments A1–A3, according to §6 of the lab manual (2025 edition) and the lab videos (which have no audio, so the content was read from the on-screen captions), is as follows.

Table 1. Procedure of basic experiments A1–A3 and the steps in the simulatorThe values are from lab manual §6.1–6.5 and the lab videos. “(provisional)” marks values that appear in neither the manual nor the videos
VideoProcedure (key points of the lab manual and the video captions)Step in simulator tab ① and default values
A1Weigh Y2O3, BaCO3 and CuO on an electronic balance (“prescribed amounts”)① Weighing: Y:Ba:Cu = 1:2:3, 0.454 / 1.587 / 0.959 g (total 3.000 g, manual §6.1)
A1First mix lightly in the dry state, then add ethanol to the wall of the mortar and keep mixing until it evaporates (liquid → slurry → paste → dry powder)② Dry mixing for 3 min (provisional; the dry premixing appears only in the video), ③ about 20 min with 1.0 mL of ethanol (provisional) (manual: do not add more ethanol after 15 min)
A1Dry for 30 min or more in a 60 °C dryer → crush lightly → put in an alumina crucible and weigh④ 30 min at 60 °C (manual and video)
A1Fire in an electric furnace at 900 °C for 10 h (done by the staff)⑤ Calcination: 3 °C/min, 900 °C, 10 h (manual). In air (specified by the instructor in charge). Furnace cooling is provisional
A2Add a few drops of ethanol to the calcined powder and grind for 30 min or more⑥ Regrinding for 30 min (wet). About 0.5 g of calcined powder is submitted for XRD
A2Put powder in the die and tap lightly, filling until the shoulder of the upper punch rises 1–2 mm. Press slowly and hold for 1 min, then unload slowly. Using the ejector stand, push down only the die to take the pellet out, and measure its mass and thickness⑦ φ 10 mm, 98.1 MPa, 1 min hold (manual §6.2; the hydraulic gauge reads about 12.0 MPa). 0.60 g of powder is provisional
A2Sintering (done by the staff; no scene in the video)⑧ 3 °C/min, 920 °C, 10 h ⑨ hold at 400 °C for 10 h → furnace cooling (manual §6.2). In air (specified by the instructor in charge)
A3Immediately after firing, measure the mass W, diameter D and thickness h; remove the as-fired surface with #400 and polish to a mirror finish with #800 (moving in a figure eight), then measure W, D and h again⑩ Polishing (0.1 mm removed in total from top and bottom is provisional) and bulk density
A3Place the polished sintered body in a polystyrene foam container, and the TA pours in liquid nitrogen (ZFC). Gently place a magnet on the cooled sintered body with bamboo tweezers (the magnet moves as if deflected away) → force the magnet down onto it, then pick it up with the tweezers and move it sideways (the sintered body follows because of flux pinning) → return to room temperature, place the magnet on the sample, and then pour in liquid nitrogen (FC) → when the magnet is moved, the sintered body follows (lab manual §6.4; the magnet is not cooled). Lab video A3 uses the arrangement with the sample placed on top of the magnet (the magnet is cooled in the same container); the order and the results are the same② Order of operations in the levitation tab (the default is the arrangement of the lab manual; the magnet is a neodymium magnet, φ 6 × 3 mm, provisional)
A3Attach four electrodes to a strip-shaped sample with silver paste (current I through the outer two, voltage V across the inner two), pass 30 mA, lower the Dewar flask with a jack to warm the sample, and record the voltage and temperature⑪ R–T by the four-probe method (model); the R–T figure in the comparison tab ③ (schematic)

The aim of this teaching material is to understand how synthesis, sintering, oxygen content and microstructure are connected to superconductivity and levitation (§5.5 of the production instructions). The simulator is the synthesis model of VEL (thermal_ybco.js, properties.js) moved onto a single-file screen for classroom use, with the following added, which VEL did not have: schematic diagrams of the interior of the material, both the arrangement with a magnet on top of the sintered body (lab manual) and the arrangement with the sintered body on top of a magnet (video), the histories of zero-field cooling and field cooling, and a comparison with superconductors other than YBCO. VEL is designed not to show students the “true state” (phase fractions and δ), but this material gives priority to showing what is happening inside (item 7 of the items to confirm in §12).

2. Raw Materials and Weighing — Reaction Equation and Mass Balance

The overall reaction can be written in a form in which CO2 leaves the carbonate and oxygen is taken up at the end.

12Y2O3 +2BaCO3 +3CuO +(14−δ2)O2 →YBa2Cu3O7−δ +2CO2 (1)

The numbers of oxygen atoms balance: on the left, 1.5 + 6 + 3 + (1/2 − δ) = 11 − δ; on the right, (7 − δ) + 4 = 11 − δ. The molar masses are 225.809 for Y2O3, 197.335 for BaCO3, 79.545 for CuO, 666.191 − 15.999δ for YBa2Cu3O7−δ, and 44.009 g/mol for CO2 (calculated from the table of atomic weights). The raw materials for 1 mol of YBCO weigh 0.5 × 225.809 + 2 × 197.335 + 3 × 79.545 = 746.21 g, so, writing m0 for the total mass of the raw materials, the mass of each reagent is as follows.

Table 2. Weighed amounts for a total raw-material mass of 3.000 gThe values in lab manual §6.1 are 0.454, 1.587 and 0.959 g, to 1 mg. The simulator calculates to 0.1 mg. A weighing error in BaCO3 (± 10 %) shifts the composition by the corresponding amount
ReagentMolar ratioMass (g)
Y2O310.4539
BaCO341.5867
CuO60.9594

The mass balance is an identity: the mass of the solid m equals the weighed mass minus the part that left as CO2, plus the net gain or loss of O2.

m=m0−mCO2+mO2 (2)

The simulator records this in a ledger at every step of the reaction and displays it as the “balance difference” in the readouts (even in a check run with node, the difference is 10−15 g or less). In pressing, each term of the ledger is scaled down by the same factor as the fraction of powder used, and in the polishing of ⑩ by the same factor as the fraction removed. When all of the BaCO3 has decomposed, the CO2 released is 2 × 44.009 / 746.21 = 11.80 % of the raw materials (by mass), and the gain or loss of oxygen (the change in δ) is added to this. Under the representative conditions, 0.3539 g of CO2 leaves after calcination and 0.0219 g of O2 enters, so 3.000 g becomes 2.668 g. Counting the theoretical yield by the method of lab manual §6.1, for the weighed values 0.454 / 1.587 / 0.959 g we have Y:Ba:Cu = 1:2.000:2.998, so Cu is very slightly the limiting reagent, and the yield is 2.645 g as YBa2Cu3O6.5 (the composition counted from the oxygen of the raw materials alone), or 2.677 g if taken as YBa2Cu3O7. The theoretical mass loss and theoretical yield are shown in the “mass balance” readout of ① (they change if a BaCO3 weighing error is entered).

3. Course of the Solid-State Reaction — Decarbonation, Green Phase, Main Reaction, Peritectic Decomposition

Simply mixing the powders does not make them react. On heating, atoms diffuse at the contact points between particles, and YBCO forms through the following stages (the reaction scheme of VEL thermal_ybco.js, carried over as it is).

  1. Decarbonation (from about 600 °C): BaCO3 + CuO → BaCuO2 + CO2. This CO2 is why the mass of the sample decreases.
  2. Green phase (from about 780 °C): Y2O3 + BaCuO2 → Y2BaCuO5 (Y211). Only when the charge is short of Ba (Ba/Y < 1.95) does Y2O3 + 2 CuO → Y2Cu2O5 (the blue phase) also form.
  3. Main reaction (diffusion-limited): Y211 + 3 BaCuO2 + 2 CuO → 2 YBa2Cu3O6.5. The layer of product separates the reactants, so diffusion through the layer sets the rate.
  4. Peritectic decomposition (above about 1015 °C in air): 2 YBCO → Y211 + liquid (3 BaCuO2 + 2 CuO) + O2. It occurs if the sintering temperature is raised too high.

The rates are first-order for decarbonation and the green phase, and follow the Ginstling–Brounshtein (GB) law for the main reaction. α is “the fraction of Y that has become YBCO.”

dξdt=kn, k=Aexp⁡(−EaRT) (3)
1−23α−(1−α)2/3 =2Dr2t, D=D0exp⁡(−EaRT) (4)

r is the particle radius, and t ∝ r2 (a powder of half the size reacts 4 times faster) is the reason mixing and grinding are done carefully (for the solid-state reaction and the Jander equation, see sintering-derivation-en.html §2). The mixing uniformity h (0–1) multiplies the rate by 0.25 + 0.75h2 and caps the attainable α at 0.62 + 0.40h (with poor mixing, Y211 and BaCuO2 remain).

Table 3. Reaction constants (values from VEL thermal_ybco.js; all are “estimates,” not fits to measurements)
StageEquationConstantsProvenance
DecarbonationFirst-order. Faster for finer powder (× √(2.5 µm/r), 0.4–2.5), × 0.5 inside a pressed pelletA = 2.32×106 s−1, Ea = 210 kJ/molEstimate
Green phaseFirst-order × (0.25 + 0.75h2)A = 8.20×107 s−1, Ea = 250 kJ/molEstimate
Main reactionGB law (4)D0 = 1.082 cm2/s, Ea = 300 kJ/molEstimate (chosen so that particles of radius 2.5 µm reach α = 0.95 at 950 °C in 12 h)
Peritectic decompositionTper = 1015 + 33 log10(pO2/0.21) °C, k = 1.67×10−3(1 − e−ΔT/8) s−11015 °C in air, 1037 °C in 1 atm O21015 °C is from Aselage & Keefer 1988 (Ref. 5); the rate is an estimate

The representative conditions in the literature are calcination at 900–950 °C for 8–12 h (Ref. 6) and sintering at around 950 °C for 6–24 h (Refs. 6, 7). It is known that partial melting begins at 940 °C when CuO coexists (Ref. 5), but this is not included in the model.

4. Mixing and Grinding — Wet Mixing in Ethanol Is a Pretreatment

Mixing in a mortar is an operation that makes the particles finer (radius r) and brings the three kinds of particles closer together (mixing uniformity h); YBCO does not form the instant you mix. The model uses the regrind of VEL as it is (t is in minutes).

r=rfloor+(r0−rfloor)e−t/τR, h=1−(1−h0)e−t/τH (5)

For wet mixing (with ethanol), rfloor = 0.6 µm, τR = 16 min and τH = 8 min; for dry mixing, 0.7 µm, 22 min and 20 min (VEL estimates). Before mixing, the particle radius is taken to be 3.0 µm and the uniformity 0.30, as if the powders had only been combined on weighing paper (the values of the VEL preset), and the ethanol evaporates at 0.055 mL/min during mixing (VEL mortar_bench.js). The boundaries between the states (liquid → slurry → paste → dry powder) are a schematic judgment based on the ratio of the volume of ethanol to the volume of powder. Once the remaining ethanol falls below 0.05 mL, the mixing is treated as dry (for 1.0 mL, the last 2.7 min or so of the 20 min). The time of switching from wet to dry is computed, so the same sample results whether you advance one step at a time on the screen or compute everything at once with a preset or the “standard procedure” of ②③. Under the representative conditions (3 min dry → 20 min wet), the uniformity goes 0.30 → 0.40 → 0.94 and the particle radius 3.0 → 1.24 µm.

In a powder that has been calcined, the reacted and unreacted portions have turned into hard lumps, so compositional unevenness should be hard to remove by regrinding. Therefore, in regrinding a calcined powder, the time constant of the uniformity τH is lengthened by up to a factor of 3 (increased smoothly from 650 °C to 850 °C; a schematic of our own, with a provisional coefficient). If mixing is omitted (0 min dry, 0 min wet), the uniformity remains at about 0.80 even after 30 min of regrinding, and even after sintering YBCO is only about 94 % (the rest is Y2BaCuO5 and BaCuO2). Because of this, for the VEL “poor mixing” condition (5 min dry, then 20 min dry after calcination), this simulator gives 86.5 % compared with 94.1 % in VEL (a deliberate difference).

5. Different Purposes of Calcination and Sintering

Table 4. Calcination and sintering
Calcination (⑤)Sintering (⑧)
PurposeReact the raw materials to make a powder of the target phase (YBCO). Drive off CO2Make the particles of the pressed pellet bond into a dense sintered body (growth of necks, reduction of pores)
Form of the samplePowder in a crucible (crushed afterward)Uniaxially pressed pellet
What happens in the materialDecarbonation, green phase, diffusion-limited main reaction (a layer of product grows from the contact points of the particles)Completion of the remaining reaction, densification and grain growth, release of oxygen at high temperature (δ increases)
If overdoneA liquid phase forms and the particles coarsen (easier to avoid by raising the temperature quickly to 900 °C: Ref. 7)Peritectic decomposition (above 1015 °C); coarsening of grains and closing of pores make it hard for oxygen to enter

Regrinding after calcination is an operation that breaks up the lumps formed in calcination, makes the particles finer, and exposes the unreacted parts as fresh contact surfaces.

If calcination is skipped: the powder is pressed with BaCO3 still in it, so decarbonation (Eq. (1)) takes place inside the pellet while the temperature rises for sintering, and CO2 comes out of the pellet. The simulator records the fraction fCO2 of the theoretical amount of CO2 that has been released inside the pellet, and includes the following schematics (all of our own, with provisional coefficients; VEL does not have them).

A sample calcined for 0 h, even with 100 % YBCO and δ = 0.098, has a relative density of 79.0 % (standard 81.7 %) and a grain-boundary connectivity of 0.40, and its resistance does not reach zero even when cooled to 40 K. The dimensions just after firing are φ 8.66 × 1.80 mm (standard φ 8.96 × 1.84 mm), and the mass is 0.534 g (standard 0.601 g). Levitation is determined by the currents inside the grains, so it still levitates, but the levitation height with field cooling falls to 1.8 mm (standard 4.0 mm) (in the arrangement with the sample on top of the magnet, magnet φ 20 × 10 mm; in the default arrangement of the lab manual, a φ 6 × 3 mm magnet levitates 0.71 mm (standard 1.29 mm)). Insufficient calcination such as 800 °C for 2 h produces the same effect in proportion to the BaCO3 that remains (18 % in the preset “Under-calcination and under-sintering”).

6. Pressing and Sintering — Densification and Grain Growth

The relative density of the green compact is assumed to rise in proportion to the logarithm of the pressure P on the sample (greenDensity of VEL sinter.js; an estimate).

ρgreen=0.60+0.185log10⁡P98.1 MPa+0.030log10⁡r1.5 µm (6)

(Limited to 0.30–0.78. Multiplied by 0.997 for a 1-min hold.) The ram of the press in the lab manual is 2.86 cm in diameter, so the hydraulic gauge reading when 98.1 MPa is applied to the sample in the die (φ 1.000 cm) is 98.1 × (1.000/2.86)2 = 11.99 ≈ 12.0 MPa (shown in the figure for step ⑦ in ①; the gauge reading is the oil pressure, not the pressure on the sample). Sintering uses the same equations as the Coble intermediate-stage model borrowed in sintering-simulator-en.html (sintering-derivation-en.html §13), and the YBCO coefficients are the VEL estimates A = 1.03×1016, Qd = 350 kJ/mol, B = 5.73×1012, Qg = 260 kJ/mol, a pore breakaway grain size of 3 µm, and ρmax = 0.96.

dρdt=Ae−Qd/RTTG4(ρmax−ρ)φ, dGdt=Be−Qg/RTTG2ψ (7)

The relative density is referenced to the theoretical density of YBCO (a first-order approximation in δ for the lattice constants, ā = 3.8571 − 0.01092(x − 6.5) Å, c = 11.7422 − 0.15325(x − 6.5) Å, x = 7 − δ, giving 6.39 g/cm3 at δ = 0 and 6.12 g/cm3 at δ = 1). The outer shape is always adjusted so that the relative density equals “mass ÷ (outer volume × theoretical density at the current δ),” so the dimensions, mass, theoretical density and relative density in the readouts are all consistent with one another. When CO2 escapes, the relative density is lowered without changing the outer shape; for the gain or loss of oxygen, the relative density (the volume fraction of solid) is kept and the outer shape is stretched or shrunk to follow the change in the volume of solid (mass ÷ theoretical density). VEL changes the relative density in proportion to the mass even for the gain or loss of oxygen, so oxygenation makes the sample appear to become denser. This is not done here, so with the same regrinding as VEL (dry, 20 min) the standard relative density comes out about 1 % lower than in VEL (a deliberate difference). With the default regrinding (wet, 30 min) the powder is finer and densification goes further, giving 81.7 %, about 5 % higher than in VEL (Appendix A.5). Difference in the reference for relative density: the lab manual uses as the denominator a constant true density derived from the lattice constants at δ = 0.49 and the composition O6.5. The denominator in this simulator is the theoretical density at the current δ, so for a sintered body with δ ≈ 0.1 the denominator differs by about 2 %, and the relative density differs from that of the manual’s method by the same amount (the manual’s true density is 6.2517 g/cm3, obtained by dividing the formula weight of YBa2Cu3O6.5, 658.19 g/mol, by the volume of the lattice at δ = 0.49 in Table 6-1 of the manual, 174.8262 Å3, and by NA. For the polished sintered body of the representative conditions (bulk density 5.20 g/cm3), the relative density is 81.7 % in the model and 83.1 % by the manual’s method. Both are shown in the readouts of ①). The sintered body of the representative conditions has a relative density of 81.7 %, a grain radius of 2.1 µm, and, just after firing, dimensions of φ 9.0 mm × 1.8 mm and a mass of 0.60 g. In the polishing of ⑩, 0.1 mm in total (a provisional value) is removed from top and bottom, and only the thickness and mass are reduced without changing the relative density (after polishing, φ 9.0 mm × 1.7 mm and 0.57 g; the bulk density W/(πD2h/4) is the same 5.20 g/cm3 before and after polishing). The “YBCO sintered body” in ②③ is the sample after polishing. The “grain-boundary connectivity” (0–1) in the readouts has a value only for a sintered body that has been heated to 800 °C or above after pressing; the YBCO fraction and the carbonate factor of the grain boundaries are multiplied by a smooth step, 3u2 − 2u3 (u = (ρ − 0.6)/0.2), that is 0 at a relative density of 60 % and 1 at 80 % (schematic). Powder and green compacts have no necks, so 0 is displayed. It is 1.00 for the standard sintered body, 0 for sintering at 850 °C for 0 h (relative density 60 %), and 0.57 for the preset “Under-calcination and under-sintering” (71.9 %). This value affects the grain-boundary step of the resistance (§10) and does not enter the levitation force.

7. Oxygen Content δ — Structure, Tc, and Speed of Oxygenation

7.1 Equilibrium oxygen content

The oxygen in the Cu–O chains of YBCO goes in and out depending on temperature and oxygen partial pressure. At high temperature it leaves and δ becomes large (tetragonal); at low temperature it takes up oxygen and δ becomes small (orthorhombic). The equilibrium oxygen content is expressed by a sigmoid (VEL’s estimate; its shape was matched to the curves of Specht et al. and is not a fit to a table).

xeq=7−δeq=6.06+0.921+exp⁡[(T−T1/2)/105], T1/2=700+65log10⁡pO2 (8)

(T is in °C and pO2 in atm.) The equilibrium δ in air is 0.094 at 400 °C, 0.133 at 450 °C, 0.360 at 600 °C, 0.575 at 700 °C and 0.871 at 920 °C, and 0.098 at 450 °C in 1 atm O2 (Fig. 1). In the literature, there is an order–disorder transition from tetragonal to orthorhombic at about 700 °C in pure oxygen (Jorgensen et al. 1987, Ref. 8), and oxygenation at 450 °C in flowing O2 gives Tc = 91 K and an oxygen content of 6.94 (Heber et al. 1987, Ref. 7).

Comparison with the figure in the lab manual: the lab manual has a figure of isotherms of δ versus oxygen partial pressure (quoted from Ref. 21). Compared with its air line, at 500–900 °C the equilibrium δ of the model is larger by 0.05–0.15, so the model leans toward the side where oxygen leaves more easily. The readings of the air line in the figure (by eye, about ±0.02) are 0.12 at 500 °C, 0.26 at 600 °C, 0.46 at 700 °C and 0.60 at 800 °C, while Eq. (8) of the model gives 0.190, 0.360, 0.575 and 0.754, respectively (at 400 °C the figure gives 0.085 and the model 0.094, nearly in agreement). For this reason, the δ at which the exchange of oxygen stops (freezes) during cooling may in reality be smaller than in the model (“furnace cooling without a hold” in §7.2).

The high-temperature crystal system is schematic: the model fixes the boundary of the orthorhombic phase at the room-temperature value x = 6.40 at all temperatures. As a result, on the equilibrium curve of Eq. (8), the switch between tetragonal and orthorhombic occurs at about 712 °C in air and about 756 °C in 1 atm O2, which is higher than in the literature (about 700 °C in pure O2, and lower in air). This is because at high temperature the boundary moves to around x ≈ 6.5. Also, at δ = 0.60–0.65 (x = 6.35–6.40), Tc of 0–10 K remains even though the structure is tetragonal (a mismatch between the Tc vanishing point of VEL, 6.35, and the lattice boundary, 6.40). Neither affects the judgment of levitation at 77 K.

Fig. 1. Equilibrium δ (Eq. (8)). The lower the oxygenation temperature, the smaller δ becomes, but the slower oxygen enters (§7.2) (our own)

7.2 Speed of oxygen uptake — diffusion distance set by the microstructure

δ approaches its equilibrium value as a first-order process, and the time constant is set by the slowest mode of the diffusion equation.

τ=Leff2π2D, D=D0exp⁡(−EaRT) (9)

Ea = 93.6 kJ/mol (0.97 eV) is the value of oxygen tracer diffusion by Rothman et al. (Ref. 9). The pre-exponential factor D0 = 6.0×10−6 cm2/s is a calibration value of VEL (about 1/23 of the value in the ab plane of a single crystal; adjusted so that the relaxation time of porous ceramics at 450 °C is 1–2 h), and must not be cited as a measured value. The diffusion distance Leff is taken to be the grain radius for open pores (relative density ≲ 0.90) and half the thickness of the pellet for closed pores (≳ 0.96), with a smooth logarithmic interpolation between them (a modeling assumption). For a grain radius of 2 µm, τ = 3.4 h at 400 °C and 1.1 h at 450 °C. That is why a hold of “400 °C, 10 h” is effective, and why, if the hold is omitted and the sample is furnace-cooled, oxygen exchange freezes near 500 °C and δ ≈ 0.19 (the equilibrium δ in the lab manual figure is smaller than the model in this temperature range, so the actual frozen δ can be smaller than this; §7.1). This δ ≈ 0.19 lies in the range where Tc changes steeply with δ (§7.3), and doubling D0 gives a δ of 0.151 (a Tc of 89.8 K), so a sample that does not levitate in ② is replaced by one that does. “It does not levitate without the 400 °C hold” is a result under the calibration value of this model (Table A2). If overfiring makes the grains grow to 5.8 µm, δ remains at about 0.30 even with the same hold. Effective diffusion coefficients in the literature scatter over orders of magnitude, and a value of 1.7×10−7 cm2/s (400 °C) has even been reported for a 94 % dense polycrystal (Ref. 10). It differs greatly from the value in the model (3.3×10−13 cm2/s at 400 °C), so the absolute value of τ is only a rough guide. The oxygen uniformity u = 1 − 0.35 log10(Leff/3 µm) (0.25–1) affects the superconducting volume fraction and the width of the transition (VEL; a modeling assumption).

7.3 Tc and δ

Tc(δ) is the piecewise-linear function of VEL (an estimate, ±3 K), with a 90 K plateau (x ≥ 6.85), a 60 K plateau (6.55–6.75), and 0 for x ≤ 6.35 (δ ≥ 0.65). Literature values differ greatly even at the same δ, depending on how the sample was made, quenching, and aging at room temperature (Table 5, Fig. 2). Superconductivity disappears at δ ≈ 0.65, near the tetragonal–orthorhombic transition (Jorgensen et al. 1990, Ref. 11; Liang et al. 2006, Ref. 12). The lab manual states that “superconductivity appears for δ < 0.6, and Tc is 90 K or higher for δ < 0.15.” The 90 K boundary agrees with the model (90 K at δ = 0.15), and the δ at which superconductivity disappears differs by 0.05 between the manual (0.6) and the model (0.65) (within the scatter of the vanishing points in the literature).

Table 5. Literature values of the oxygen content and Tc (compiled values from Table 1.1 of Vázquez-Navarro 1998, Ref. 10) and the model
7 − δδJacobson et al. (K)Poulsen et al. (K)Cava et al. (K)Model (K)
6.990.0194939491.9
6.970.0392939291.6
6.910.0984.3919090.8
6.870.136386.58990.3
6.700.306066.582.959.5
6.620.3857.16168.657.1
6.540.4650586051.5
6.450.5547.154.257.620.0
6.400.603147.957.510.0

Fig. 2. Tc and δ. The line is the model of the simulator and the symbols are the literature values in Table 5 (our own)

The model gives values lower than the literature at δ = 0.55–0.6 (a steep drop to match the vanishing point 6.35 of the VEL formula). What matters for the levitation result is the neighborhood of where Tc exceeds 77 K (δ ≈ 0.15–0.2), and this lies within the range of the literature. The superconducting volume fraction is given, using the transition width w = 1 + 20 max(0, 0.85 − ρ) + 40(1 − f) + 24(1 − u) K (f: mass fraction of YBCO), by fsc(T) = f u / [1 + exp(−(Tc − T)/(0.45w))] (VEL properties.js; an estimate). The oxygen occupation of the Cu–O chains in the internal figure is 1 − δ divided between the chain sites (b direction) and the sites that were vacant (a direction) by the order parameter s = o(x)/o(7) (n1 = (1 − δ)(1 + s)/2, n5 = (1 − δ)(1 − s)/2, o(x) = 0.0170√((x − 6.40)/0.60), tetragonal for x ≤ 6.40).

8. Superconductivity and Levitation — Meissner Effect and Flux Pinning

8.1 Meissner state and mixed state

A superconductor expels a weak magnetic field from its interior by shielding currents at the surface (the Meissner effect). YBCO is a type II superconductor, and when the field exceeds the lower critical field Hc1 (a few mT to a dozen or so mT at 77 K), it enters the mixed state, in which magnetic flux penetrates as quantized flux (vortices) of Φ0 = h/2e = 2.07×10−15 Wb each (for the distinction between type I and type II, see superconductivity-simulator-en.html ①). The surface of the NdFeB magnet in the levitation experiment is at 0.1–0.5 T, so the inside of the sintered body is in the mixed state. Most of the flux has entered the sample, and only a part is expelled.

The flux that has entered gets caught on second phases (such as Y2BaCuO5), defects and grain boundaries and cannot move (flux pinning). Because the flux cannot move, a shielding current flows when you try to change the magnetic field, and its upper limit is the critical current density Jc. With perfect diamagnetism alone, the lateral position above a flat superconductor is unstable, and the penetration of flux and pinning make the position stable (Hellman et al. 1988, Ref. 13). It is not that “it floats because the resistance is zero”; the magnitude of the shielding current and the force that holds the flux in place determine how it floats.

8.2 Zero-field cooling (ZFC) and field cooling (FC)

A note on terminology The background section of the lab manual does not call the diamagnetism seen in ZFC the Meissner effect, because it can also be derived from perfect conductivity, and instead calls the expulsion of flux when the sample goes below Tc in FC the Meissner effect. The distinction made in this page and the simulator (repulsion by shielding currents in ZFC; the Meissner effect is the part of the flux pushed out in FC) is the same as this. On the other hand, the procedure section of the manual and the captions of the lab video call the repulsion in ZFC the “Meissner effect” (an inconsistency in usage).

8.3 Magnetic field of the magnet

The magnet is a uniformly magnetized cylinder (remanent flux density Br, radius a, thickness b), and only the field on the central axis is used (z is the distance from the face of the magnet).

Bz(z)=Br2 [z+b(z+b)2+a2−zz2+a2] (10)

The default magnet is a neodymium magnet φ 6 mm × 3 mm matched to the arrangement of lab manual §6.4 (a magnet placed on the cooled sintered body), with Br = 1.25 T (provisional; the manual gives no material or dimensions; §12), a surface-center field of 0.44 T and a mass of 0.64 g (weight 6.2 mN). In this model only a small magnet levitates when placed on the sintered body (with field cooling, 1.29 mm for φ 6 × 3 mm and 0.05 mm for φ 8 × 3 mm), and the size can be changed with a slider. A magnet with a diameter larger than the sample (φ 8.95 mm for the standard sintered body), that is, φ 10 mm or more, does not levitate regardless of its thickness at the default Br = 1.25 T. The effective area Aeff in Eq. (12) is set by the smaller cross-section of the sample and the magnet, so the part sticking out beyond the sample does not contribute to the repulsion (and the larger the diameter, the gentler the field gradient on the axis), while only the weight of the magnet increases as the square of the diameter. Compared at a thickness of 3 mm, the repulsion on contact (ZFC) falls once the diameter exceeds the sample: 19.7 mN for φ 8, 20.0 mN for φ 9, 16.4 mN for φ 10, 11.2 mN for φ 12 and 3.2 mN for φ 20 (for φ 10 it falls below the weight of 17.3 mN). Although outside the range of the slider, φ 10 × 1 mm is lighter than the default magnet (weight 5.8 mN) but gives a repulsion of only 4.7 mN and does not levitate. Conversely, φ 8 × 5 mm is heavy (weight 18.5 mN), but its repulsion is 26.7 mN, so it levitates by 1.66 mm in ZFC. For the arrangement “sample placed on top of the magnet” (for reference: lab video A3), a φ 20 mm × 10 mm magnet (provisional; the ratio of diameter to thickness is the same, so the surface-center field is the same 0.44 T) is used, matched to the appearance in the video (about twice the diameter of the sample, and a cylinder thicker than the sample).

8.4 Force model (a schematic for teaching, of our own)

The diamagnetic magnetization of the sample m (A/m, positive when opposite to the field) is expressed as the sum of a reversible part and an irreversible part due to pinning.

m=fsc(T) [Hp(T)HH2+Hp2η +Mcφ], Mc=Jc(T,B)ℓ3 (11)

The fsc of a sintered body is the superconducting volume fraction of the model in ① (mass fraction of YBCO × oxygen uniformity × transition sigmoid) multiplied by a porosity coefficient (ρrel/0.85)1.5 (capped at 1). It is a coefficient expressing that the superconducting volume decreases by the amount of the pores, and that mutual shielding between grains becomes weaker when contact between grains is small; the factor of meissnerTest in VEL properties.js was borrowed (a modeling assumption). Without it, an unsintered porous sample would repel more strongly than a dense one. For the sample sintered at 850 °C for 0 h (relative density 60 %) the coefficient is 0.59, and the ZFC force on contact is 8.4 mN (standard: 12.3 mN, when placed on a φ 20 × 10 mm magnet).

The force is obtained by integrating over the thickness of the sample the force with which the diamagnetic magnetization is pushed toward the weaker field (Aeff = π min(Rs, a)2).

F(z)=Aeff∫zz+tm(z′)|∂Bz∂z|dz′ ,F(zeq)=W (12)

The levitation height zeq is the point where F crosses the weight W from above to below, searching upward from contact (z = 0) (W is the weight of the sample in the arrangement with the sample on top of the magnet, and that of the magnet in the arrangement with the magnet on top of the sample; the density of NdFeB is 7.5 g/cm3). If F(0) < W, it does not levitate. Lateral stability is judged by the ratio S of the restoring force of pinning, kpin ∝ Mc χ (χ: 1 for FC and after pressing down, 0.25 for ZFC), to kdia ∝ 0.15 × the first term, the part of the reversible diamagnetism that pushes outward from the center (stable at 1.2 or more, slides off below 0.8). This lateral judgment is a schematic that expresses only the tendency “the stronger the pinning, the more stable,” and does not calculate the lateral force due to the shape of the magnet. The coefficients 0.15 and χ = 0.25 are provisional values, and whether a sintered body placed in ZFC slides off is determined by how these coefficients are chosen. In fact, in ZFC of a sintered body, S ≈ 0.5 regardless of the dimensions, Br and arrangement of the magnet (for magnets of φ 6–20 mm, and also for a ferrite magnet). S is also proportional to Mc ∝ Jcℓ, so it changes greatly with the intragrain Jc (provisional value 105 A/cm2) and the grain radius (about 2 µm). How a sintered body levitates also depends strongly on the magnitude of the reversible diamagnetism Hp0 (10 mT, provisional) and on the fraction of flux pushed out in FC, pexp (1, provisional). Of the 12.3 mN force of the standard sintered body (77 K, ZFC, contact, magnet φ 20 × 10 mm in the arrangement with the sample on top of the magnet), the reversible diamagnetism (first term) accounts for 9.6 mN (78 %) and pinning (second term, the term that saturates at the intragrain critical current) for 2.7 mN (22 %) (the breakdown is shown in the readouts of ②). With Hp0 set to 5 mT in the same arrangement, it does not levitate in FC, and in ZFC it stops while wobbling at S = 0.97. With 20 mT, it levitates 10.0 mm in FC, and in ZFC it slides off at S = 0.25. With pexp = 0.5, it does not levitate in FC; with the intragrain Jc tripled, it is stable in ZFC (S = 1.52, 10.7 mm) and hardly levitates in FC. In the default arrangement of the lab manual (a φ 6 × 3 mm magnet on the sintered body), with Hp0 = 5 mT the magnet levitates 1.54 mm in ZFC and wobbles at S = 0.99, and hardly levitates after pressing down or in FC (0.13 mm, 0.01 mm). With 20 mT it levitates 2.99 mm after pressing down and in FC, but wobbles at S = 1.02, and the sintered body follows with a delay. With pexp = 0.5 it hardly levitates in FC (0.01 mm), and with the intragrain Jc tripled, S = 1.53 in ZFC and the magnet is not deflected (the opposite of [2] in the manual). Therefore, [2] of lab manual §6.4 (the magnet is deflected), [3] and [4] (the sintered body follows when the magnet is moved), and the “slides off in ZFC” and “levitates in FC” of video A3 can all come out the other way in either arrangement, depending not only on 0.15 and χ but also on the choice of Hp0, pexp and Jc (all provisional) (item 9 of the items to confirm in §12). In the arrangement with the sample on top of the magnet (φ 20 × 10 mm), after the sintered body has been pressed down, the flux pushed in is not strong enough to hold the position (at the height where it stops, Mc ≈ 600 A/m, which is about 29 % of the reversible magnetization mrev ≈ 2100 A/m of the first term of Eq. (11) that produces the repulsion). In the direction away from the pressed position, however, the second term of Eq. (11) reverses to pull back (−2.27 mN, canceling about 29 % of the 7.84 mN repulsion of the first term), and the sample stops at a height lower than when it is gently placed in ZFC (8.3 mm), almost the same as in field cooling (4.0 mm) (6.6 mm with the first term alone). The same holds in the default arrangement (a φ 6 × 3 mm magnet on the sintered body): at the height where it stops, Mc ≈ 620 A/m and the second term is −2.60 mN (about 29 % of the 8.84 mN repulsion of the first term), and the magnet stops at 1.29 mm, lower than in ZFC (2.50 mm) (2.00 mm with the first term alone). The effect of pressing down is not just that χ changes from 0.25 to 1. When a melt-grown bulk is pressed down, part of the flux pushed in remains, and it stops at 16.0 mm, lower than when it is gently placed in ZFC (20.3 mm) (hysteresis; magnet φ 20 × 10 mm). In the default lab manual arrangement (a φ 6 × 3 mm magnet placed on the sintered body), the ZFC force at contact for the standard sintered body is 17.2 mN (first term 13.4 mN, second term 3.8 mN), which exceeds the 6.2 mN weight of the magnet, so the magnet levitates 2.50 mm, but at S = 0.50 it is deflected and slides off ([2] of the manual: “the magnet moves as if deflected away”). When it is pressed down it stops at 1.29 mm (S = 1.97), and with field cooling it is also 1.29 mm (S = 1.97); in both cases the sintered body follows when the magnet is moved sideways ([3] and [4] of the manual).

Table 6. Values put into the levitation model (all are provisional values for teaching; the basis is the order-of-magnitude range)
MaterialJc77 (A/cm2)ℓμ0Hp0pexpBirr(77 K)Basis
YBCO sintered body1×105 (intragrain)Grain radius (result of ①, about 2 µm)10 mT17 TThe intragrain Jc is of the order of 104–106 A/cm2 (Refs. 14, 15)
YBCO melt-grown3×104Sample radius10 mT0.17 TJc(77 K, 1 T) > 104 A/cm2 (Ref. 15)
GdBCO (TSIG)4×104Sample radius10 mT0.18 TOrder of magnitude of the levitation force in Wu et al. 2024 (Ref. 16)
Bi-2223 sintered body1×104 (intragrain)3 µm4 mT10.25 TIrreversibility field at 77 K of 0.2–0.3 T (Ref. 18)

8.5 Comparison with the ideal case (our own derivation)

When there is a vertical point magnetic dipole md (A·m2) at height z above an infinitely wide perfect diamagnet, the image that satisfies Bz = 0 at the boundary is an oppositely directed dipole at −z, and the repulsive force becomes as follows (method of images, our own).

FZFC=3μ0md232πz4, FFC=3μ0md22π[1(2z)4−1(z+z0)4] (13)

The FC expression was derived by ourselves from the description of Kordyuk’s frozen-image method (Ref. 17), which adds a “frozen image,” the mirror image at the cooling position z0 with its magnetization reversed, and it has not been checked against the equation in the original paper. At z = z0, F = 0; for z > z0 the force is attractive (hanging); and as z0 → ∞ it returns to the ZFC expression. Example: for a magnet of φ 10 mm × 5 mm and Br 1.2 T (md = 0.375 A·m2, 2.9 g), the ZFC equilibrium is at about 21 mm (a gap of 18 mm), which is far larger than the actual value for a sintered body (a few mm). This is because the sample is finite, the grain boundaries are weak links, and flux penetrates. Calculating a melt-grown bulk (Mc ≫ H) with Eqs. (11) and (12) approaches this ideal.

8.6 Why the force differs by orders of magnitude between a sintered body and a melt-grown bulk

  1. Weak links at grain boundaries: at randomly oriented grain boundaries Jc falls exponentially with the misorientation, so the intergrain Jc of a polycrystal is small (Hott’s review, Ref. 15).
  2. The current loops close inside the grains: since Mc ∝ Jcℓ, the difference in size between a grain (µm) and a domain (mm to cm) becomes a difference of a factor of about 103 as it is (Kordyuk’s granular model, Ref. 14).
  3. Pinning sites: a melt-grown bulk has fine particles of Y2BaCuO5 dispersed in it to strengthen the pinning.
  4. Pores, second phases and oxygen unevenness reduce the superconducting volume and the current paths.

8.7 Checking the orders of magnitude

9. Comparing Superconductors — YBCO Sintered Body, Melt-Grown Bulk, GdBCO, Bi-2223

Table 7. The superconductors compared (Tc, preparation method and microstructure are literature values; the values of the levitation model are in Table 6)In tab ③ they are compared with the same magnet (φ 12 × 5 mm, 1.25 T), the same dimensions (φ 10 × 2 mm) and the same temperature
MaterialTcPreparation and microstructureProperties relevant to levitation at 77 KModel result (77 K)Source
YBa2Cu3O7−δ sintered body90.7 K (representative conditions of ①)Solid-state reaction, randomly oriented polycrystalWeak links at the grain boundaries; granular model25 mN at 1 mm (ZFC). In ZFC it levitates but is laterally unstable; in FC it levitates 3.8 mm and is stableRefs. 14, 15
YBCO melt-grown92 KMelt growth (TSMG/MPMG), single domain, Y211 dispersedJc(77 K, 1 T) > 104 A/cm21.46 N at 1 mm. In ZFC it levitates 12.7 mm and is stable; in FC it is held strongly in contactRef. 15
GdBa2Cu3O7−δ94 KTop-seeded infiltration growth (TSIG), single domain, Gd211 dispersed14.7–35.6 N for a diameter of 20 mm (Ref. 16)1.41 N at 1 mm. Similar to melt-grown YBCO. Even at 90 K, 0.24 N remainsRefs. 15, 16
(Bi,Pb)2Sr2Ca2Cu3O10 (Bi-2223) sintered body110 KSolid-state reaction, plate-like grainsThe irreversibility field at 77 K is low, 0.2–0.3 T, so the pinning is weak (about 7 T for YBCO)12 mN at 1 mm. It levitates but is laterally unstable in both ZFC and FC. It levitates even at 90 K (7.8 mN at 1 mm > weight 7.7 mN; levitation height 1.2 mm)Refs. 18, 20

Although Bi-2223 has the highest Tc, at 77 K the flux can move in the field near the magnet (0.2–0.4 T), so even if it levitates its position is not fixed. This is a comparison that shows it is not the case that “below Tc, any material levitates at the same height” (the numerical levitation forces are schematic).

10. Temperature Dependence of Resistance by the Four-Probe Method

In the four-probe method, which separates the current terminals from the voltage terminals, the resistance of the sample can be measured without including the contact resistance (video A3, lab manual §6.5). In the lab, four electrodes are attached with silver paste to a sample cut into a strip (the outer two for the current I, the inner two for the voltage V, with L the distance between the inner terminals), a constant current of 30 mA is passed, and the sample, immersed in liquid nitrogen, is warmed by lowering the Dewar flask with a jack while the voltage and temperature are recorded from about 77 K to about 173 K (−100 °C). Step ⑪ of ① draws this measurement as R/R(300 K) from the model of the sample of ① (the absolute value of the resistivity is not treated). In a sintered body the resistance drops in two stages, the transition inside the grains (Tc) and the transition of the weak links at the grain boundaries (Tc − transition width), and with many second phases or pores the curve is one in which “the resistance starts to drop but does not reach zero” (resistivityCurve of VEL properties.js, simplified by normalizing to R/R(300 K)). The curves of the other materials are schematics that represent only the position of the transition (zero resistance of Bi-2223 at 107 K: Ref. 20). Zero resistance and levitation are separate properties; Bi-2223 has zero resistance at 77 K, but its levitation is unstable.

11. Simplifications and Range of Validity of the Model

12. Items to Confirm (with Prof. Mochizuki)

  1. The reagent masses were fixed by lab manual §6.1 (0.454 / 1.587 / 0.959 g, total 3.000 g). (Resolved)
  2. The temperatures and times of sintering, the 400 °C hold and furnace cooling were fixed by lab manual §6.2. The atmosphere is air (student experiments; in research, a tube furnace with flowing N2 or a glove box filled with Ar is also used). (Resolved: Prof. Mochizuki’s reply, 2026-09-25) The cooling after calcination (furnace cooling?) is in neither the manual nor the reply, so it remains provisional. The preset of 450 °C for 12 h in flowing oxygen is kept as a condition from the literature and for comparison.
  3. The die diameter of φ 10 mm, 98.1 MPa and a 1 min hold were fixed by lab manual §6.2. The amount of powder used for pressing (provisionally 0.60 g) and the amount removed by polishing (provisionally 0.1 mm in total from top and bottom) are not in the manual (the hydraulic gauge reading is the oil pressure, not the pressure on the sample; it is about 12.0 MPa when 98.1 MPa is applied).
  4. The material, dimensions and surface flux density of the magnet used for levitation, the dimensions and mass of the sintered body, and measured examples of the levitation height. Prof. Mochizuki’s reply (2026-09-25) is “follow the lab manual,” but §6.4 of the manual gives no material or dimensions for the magnet, so the default is a neodymium magnet of φ 6 × 3 mm and 1.25 T (provisional) (a size that levitates when placed on the sintered body; §8.3). The dimensions and mass of the sintered body and measured examples of the levitation height are unconfirmed.
  5. The wet mixing time of about 20 min (no ethanol is added after 15 min) was fixed by the lab manual. The time of the dry premixing (which appears only in the video) and the amount of ethanol are provisional.
  6. Whether comparison superconductors (GdBCO bulk, Bi-based) are available in the lab. The simulator compares them using representative literature values.
  7. The simulator calculates from the inputs and displays δ, the phase fractions, the mass balance (including the mass that left as CO2), the bulk density, the relative density, and the theoretical density (the value at the current δ; shown in the readouts). If you stop partway through the cooling of ⑨, for example, when δ is close to 0.5, the theoretical density reads almost the same as the true density obtained by the method of the lab manual. (Resolved: Prof. Mochizuki’s reply, 2026-09-25) The lab manual is that of the former Tokyo Institute of Technology and is not distributed at Tokyo University of Science, so all the calculated values may be shown. The theoretical mass loss and theoretical yield (§2) and the values at both ends of the theoretical density (§6), which had been removed, were restored, and the true density by the manual’s method, 6.2517 g/cm3, and the relative density (§6), the hydraulic gauge reading (§6), and the reading of the equilibrium δ in the manual’s figure (§7.1) were added. They are also shown in the readouts of ① and in the figure of ⑦.
  8. Tc(δ) uses the piecewise-linear function of VEL. How to treat the scatter in the literature (Table 5), and which series it should be matched to.
  9. The levitation arrangement. (Resolved: Prof. Mochizuki’s reply, 2026-09-25) The correct levitation arrangement is the “magnet on top of the sample” of lab manual §6.4, so this was made the default and the procedure was also made to follow the manual (gently place the magnet on the sintered body cooled in ZFC → force the magnet down and move it sideways → FC, in which the magnet is placed on the sample at room temperature and the sample is then cooled → move the magnet; the magnet is not cooled). The arrangement of the video (sample on top of the magnet) was kept for reference. With the default φ 6 × 3 mm magnet, in ZFC the magnet levitates 2.50 mm and then is deflected and slides off (S = 0.50); when pressed down it is 1.29 mm (S = 1.97), and in FC also 1.29 mm (S = 1.97); in both cases the sintered body follows when the magnet is moved sideways. The trends agree with [2]–[4] of the manual (the “small NdFeB magnet levitates stably 1–3 mm” that VEL had taken as a textbook measurement is not found in the 2025 edition of the manual). Remaining items to confirm: how far the magnet levitates in FC (not stated in the manual) and what ZFC looks like in the lab. The coefficients of the lateral stability (§8.4), and the provisional values of Hp0 (10 mT), pexp (1) and the intragrain Jc (105 A/cm2), which can reverse the way a sintered body levitates, should be decided to match measurements.
  10. Whether to bring the equilibrium δ equation (the VEL estimate) closer to the figure in the lab manual (in air at 500–900 °C, δ is 0.05–0.15 smaller than in the model) (§7.1).

Appendix A. What the Simulator (ybco-simulator-en.html) Shows

What this appendix explains What each tab of the companion simulator ybco-simulator-en.html draws, what each operation means, and how the numbers on the screen are calculated.

A.1 The page as a whole

There are three tabs. In every tab, the figures are on the left (from the top: the apparatus, a schematic of the interior of the material, and a graph), with the legend, the comment box and the readouts below them, and the controls on the right (below, on a smartphone). The three figures and the numbers are all drawn from a single state (the sample). Appending #synth, #levitate or #compare to the address opens that tab; ?embed=1 gives an embedded view with the heading and the introduction hidden, and something like ?only=levitate shows only that tab. The animation stops when a tab is not displayed or when the screen is hidden.

Table A1. Correspondence between the tabs and the main text
TabWhat to look at on the screenMain text
① Synthesize a sample (synth)The 11 steps from weighing to oxygenation, polishing and the four-probe method. The apparatus, particles and phases, pores, oxygen entering and leaving, oxygen in the Cu–O chains, temperature history with phase fractions, relative density and δ, Tc(δ), bulk density, R–T§2–§7, §10
② Cool and levitate (levitate)Cross section of the liquid nitrogen, magnet and sample; magnetic field lines and quantized flux; the force F(z) and the weight; time variation of the temperature and the levitation height§8
③ Compare superconductors (compare)How the four materials levitate under the same conditions, F(z), the temperature dependence of the force and the resistance, the table of materials§9–§10

A.2 ① Synthesize a sample

“Next step” advances one step at a time, and “Run all steps in a row” advances to the end, by animation. “Compute the result right away” skips the animation. The sample from the previous step is carried over as it is to the next step, and a change of conditions affects only the steps yet to start (not a step that is running or paused). The conditions of ⑨ (holding temperature, time, atmosphere and cooling) take effect even if you change them after ⑧ has finished. This is because, when ⑨ starts, the furnace program is rebuilt with the conditions at that time, starting from the point where the hold of ⑧ ended. The heading and set temperature in the furnace figure show the conditions of the program actually being used. The name of “Keep the last result” is made from the conditions actually used to make that sample (if they differ from a preset, the main conditions are listed). When you move a slider, the preset field becomes “(conditions changed from a preset).” The labels “green compact” and “sintered body” in the interior figure and the readouts are distinguished by whether the sample was heated to 800 °C or above after pressing (schematic). The sample is complete after ⑨ (the name of “Keep the last result” no longer carries “(up to …)”), and once the polishing of ⑩ is finished you can select “Sample made in ①” in ② and ③. ⑪ does not change the sample. Time is compressed in playback (at playback speed ×2, one frame is 0.4 min for mixing and 12 min for heat treatment), and the actual elapsed time is shown in the apparatus figure and the readouts. ⑦, ⑩ and ⑪ are steps that show a scene, so regardless of the playback speed they are drawn in about 8 s, 5 s and 7 s, respectively.

Table A2. Results of the presets (values calculated with node)Polished sample, 77 K. The levitation values are those obtained when the same sample is selected in tab ②. The left is the arrangement with the sample on top of the magnet (for reference: the video; magnet φ 20 × 10 mm, 1.25 T), and the right is the default arrangement of the lab manual (a φ 6 × 3 mm, 1.25 T magnet placed on the sintered body)
PresetYBCOδTcRelative densityLevitation: sample on magnet (ZFC / FC)Levitation: magnet on sample (ZFC / pressed down and FC)
Lab manual procedure (sinter at 920 °C for 10 h → 400 °C for 10 h → furnace cooling)100 %0.10190.7 K81.7 %ZFC: levitates but slides off / FC: stable at 4.0 mmZFC: levitates 2.50 mm and is deflected (S 0.50) / pressed down and FC: stable at 1.29 mm (S 1.97), the sintered body follows
Flowing O2 at 450 °C for 12 h (literature/comparison conditions)100 %0.08890.8 K81.7 %Same as aboveSame as above (2.52 mm / 1.30 mm)
No 400 °C hold (furnace cooling, in air)100 %0.19277.9 K81.6 %Does not levitate (0.6 mN at contact < weight 5.6 mN). With furnace cooling in flowing O2, δ is 0.142, Tc is 90.1 K, and it levitates. δ ≈ 0.19 lies in the range where Tc changes steeply (§7.3), and whether it levitates or not is reversed by the choice of the calibration value D0 (doubling D0 gives δ 0.151 and Tc 89.8 K, and it levitates at 3.8 mm in FC; halving it gives δ 0.246 and Tc 62.1 K). This is also stated in the comment of ②Does not levitate (0.8 mN at contact < magnet weight 6.2 mN)
Quench from the sintering temperature100 %0.86081.6 %Not superconductingNot superconducting
Over-sintering, 980 °C, 20 h100 %0.30259.5 K86.2 %Not superconducting (77 K). Grain radius 5.8 µmNot superconducting (77 K)
Over-firing, 1050 °C, 4 h40 %0.825079.9 %Peritectic decomposition. Y211 21 %, BaCuO2 32 %Not superconducting
Weigh 10 % too much BaCO377 %0.10190.7 K81.7 %ZFC: slides off / FC: hardly rises (0.04 mm; 17 % BaCuO2 remains)ZFC: deflected at 1.87 mm / pressed down and FC: 0.45 mm (S 1.92)
Under-calcination and under-sintering (800 °C, 2 h; 880 °C, 4 h)89 %0.08990.8 K71.9 %FC: 3.6 mm but wobbles (the grains are small and Mc is small). 18 % of the CO2 was released inside the pelletZFC: deflected at 1.66 mm / pressed down and FC: 0.97 mm but wobbles (S 1.04)
(Slider) calcination 0 h100 %0.09890.7 K79.0 %FC: 1.8 mm. More pores and cracks, grain-boundary connectivity 0.40, the resistance does not reach zero. The dimensions are smaller than standard (there is less solid by the amount of CO2)ZFC: deflected at 1.98 mm / pressed down and FC: 0.71 mm (S 1.85)
(Slider) sintering at 850 °C for 0 h94 %0.08390.9 K60.0 %Slides off in both ZFC (5.6 mm) and FC (4.0 mm) (S 0.17, 0.66). Sintering is insufficient (grain-boundary connectivity 0)Slides off in both ZFC (1.21 mm) and pressed down and FC (0.72 mm) (S 0.17, 0.67)
(Slider) dry 0 min, wet 0 min94 %0.09890.7 K76.2 %FC: 3.3 mm. Y211 and BaCuO2 remainZFC: deflected at 2.19 mm / pressed down and FC: 1.00 mm (S 1.85)

A.3 ② Cool and levitate

The sample is chosen from “Sample made in ①” (selectable once ① has been advanced to the polishing of ⑩), four kinds of YBCO sintered body (ready-made samples calculated with the same model as ①), and melt-grown YBCO, GdBCO and Bi-2223 (assumed to be cut out to φ 10 × 2 mm). The arrangement can be switched between “Magnet placed on top of the sample (lab manual §6.4)” (the default; neodymium magnet φ 6 × 3 mm, provisional) and “Sample placed on top of the magnet (for reference: lab video A3)” (magnet φ 20 × 10 mm, provisional). In the arrangement of the lab manual the magnet is not cooled but kept outside the container (at room temperature). In the arrangement of the video, in ZFC the magnet is also cooled in the same polystyrene foam container (this is shown in the figure “View of the experiment” and in the comment). The control panel notes that the liquid nitrogen is poured by a TA, that leather gloves (not cotton work gloves) and bamboo tweezers are to be used, and warns about frostbite and oxygen deficiency. The order of operations becomes the history: pouring liquid nitrogen away from the magnet gives ZFC; stacking the magnet and the sample at room temperature and then pouring gives FC; pressing the cooled sample and the magnet together (in the arrangement of the manual, forcing the magnet onto the sintered body) gives “pressing down.” When the sample is warmed, Jc and the superconducting volume fraction decrease and the levitation height drops, the sample loses its lateral stability or can no longer support the weight, and above Tc the history is erased. Time runs about 12 times faster (0.2 s per frame). The time constants of cooling and warming (8 s, 1500 s) are provisional values. In the “magnet on the sample” arrangement (the default), it is the magnet that levitates or is deflected, so the subject of the comment is the magnet, and the buttons become “Gently place the magnet on the cooled sintered body” and “Force the magnet down (press it in to pin the flux).” As in lab manual §6.4, the button for moving is “Move the magnet sideways (does the sintered body follow?)”, and it picks up the upper magnet with tweezers and moves it by ±7 mm. When the pinning keeps the relative position (S ≥ 1.2 after field cooling or after pressing down), the sintered body follows by sliding along the bottom of the container (on the base in the figure); if it is levitating and 0.8 ≤ S < 1.2, it follows with a delay (after the magnet returns, it keeps moving until it catches up with the original relative position); otherwise the sintered body stays where it is and only the magnet moves. When it does not follow, the distance moved is reduced to (sample radius − 1 mm) so that the center of the magnet does not go beyond the edge of the sample. Even when the magnet is just gently placed after zero-field cooling, if it is levitating and S ≥ 1.2 (a sample with strong pinning, such as melt-grown YBCO or GdBCO), the sample follows because of the pinning of the flux that entered near the surface as the magnet was brought close. The comment in this case does not say “pushed in” but says that this is a condition not in the procedure of lab manual §6.4 (with the sintered body of the manual the magnet is deflected, as in [2]). The comment after the sintered body has been pressed down is written from the calculated equilibrium height and the breakdown of the force (the first and second terms): the flux pushed in is not strong enough to hold the position, but it acts as a force pulling back in the direction away from the pressed position, so the sample stops at almost the same height as with field cooling (lower than in ZFC). The quantities compared are the pinning magnetization Mc and the reversible magnetization mrev (§8.4). When the force is still short of the weight even after pressing down, the comment says “does not levitate even when pressed down,” and the lateral stability S is displayed only while the sample is levitating. When the sample has slid off, the comment says that the lateral judgment is decided by the provisional coefficients (0.15; χ = 0.25 for ZFC) and does not depend on the dimensions or arrangement of the sample and the magnet. When it has slid off after pressing down or field cooling, no next operation is recommended. In the breakdown of the force in the readouts, when the pinning term is in the pulling-back direction (negative), no percentage is given and it is shown with its sign. In the arrangement with the sample on the magnet, if “Move the magnet sideways” is pressed for a sample that is not levitating (or is still almost in contact), the sample moves together with the magnet, remaining on top of it. If the sample is superconducting and the lateral stability S is 1.2 or more after field cooling or after pressing down, the comment says “the pinned flux keeps the relative position” (as in [3] and [4] of lab manual §6.4); otherwise it says “this is due to friction, not pinning.” The lateral judgment in the comment after pressing down is also written differently with the same thresholds (stable at S ≥ 1.2, wobbling at 0.8 or more). For a sample that stops very close to contact after being pressed down, the second term reverses in the direction away from the pressed position to pull back and cancels part of the repulsion, so the comment says that it balances the weight very close to contact (it is the pinning that holds it back), and shows Mc compared with mrev (if a magnet made φ 8 × 3 mm with the slider is pressed onto the standard sintered body, it stops at 0.15 mm, where the first term is 15.34 mN and the second term is −4.29 mN; with the first term alone it would levitate to 1.77 mm). As in the table and the readouts of ③, S in the comment is written as “10 or more” when it exceeds 10. The readouts show the breakdown of the force into reversible diamagnetism and pinning. The scales of the figure “View of the experiment” and of the flux figure, the horizontal axis of the force graph, and the axis of the levitation height change to suit the levitation height and the dimensions of the magnet (the scale of the figure “View of the experiment” is shown at its lower right). Whether the sample is superconducting is judged by T < Tc and a superconducting volume fraction above 2 % (the smooth tail of fsc remains above Tc, but the force is 0 for T ≥ Tc).

A.4 ③ Compare superconductors

The four materials are placed under the same magnet, with the same dimensions (φ 10 × 2 mm) and at the same temperature, and the balance and the lateral stability are shown side by side for ZFC (cool, then place on the magnet) or FC (place on the magnet, then cool). To make the differences between materials easy to see, ③ compares them in the arrangement with the sample on the magnet (for reference: lab video A3; the weight is that of the sample). The lab procedure (lab manual §6.4) is the default arrangement of ② (magnet on the sample). The temperature ranges from 64 to 115 K (77 K is liquid nitrogen, and temperatures below it are for reference, assuming liquid nitrogen under reduced pressure; nitrogen solidifies at its triple point of 63 K, so the lower limit was set at 64 K). Whether the comment says the sample levitates or not takes the lateral stability S into account, as in the figure and the table (if S < 0.8, it says “there is repulsion but it slides off sideways”; in FC it says “levitates” only when the sample has risen higher than 0.05 mm). The F(z) curves of melt-grown YBCO and GdBCO almost overlap, so the melt-grown one is drawn as a thick line underneath. The graphs below show the temperature dependence of the force at a gap of 1 mm (ZFC) and of the four-probe resistance R/R(300 K) (schematic). The table lists Tc, the preparation method and the microstructure (literature values), and the Jc, ℓ and Birr put into the model (provisional values). If you check “Use ‘Sample made in ①’ as the YBCO sintered body,” the comparison uses the sintered body made in ① (its δ, grain size and relative density). The “force (1 mm)” in the readouts is calculated for the cooling method currently selected (ZFC/FC), and if it is attractive it is marked “(attraction).”

A.5 Notes on the calculation and checks

References

  1. Lab manual for Materials Science Experiments (Inorganic Materials), former Tokyo Institute of Technology, 2025 edition (§5, §6.1–6.6) (in Japanese). It is not distributed in courses at Tokyo University of Science. It was used to check the procedures and values; its text and figures are not reproduced. Lab videos (A1–A3: weighing, mixing and calcination of the YBCO raw materials; grinding and pressing; polishing, observation of the Meissner effect and measurement of the electrical resistance). Materials in Prof. Mochizuki’s possession (not public). The atmosphere of the heat treatment, the levitation arrangement and the display of calculated values follow Prof. Mochizuki’s instructions (2026-09-25).
  2. VEL (virtual laboratory, in preparation) src/sim/thermal_ybco.js, properties.js, sinter.js. All constants carry the note “estimate (for teaching).”
  3. Mochizuki Research Group sintering-simulator-en.html and sintering-derivation-en.html (the equations of sintering).
  4. Mochizuki Research Group superconductivity-simulator-en.html and cooper-pair-simulator-en.html (the principles of superconductivity).
  5. T. B. Aselage, K. D. Keefer, “Liquidus relations in Y–Ba–Cu oxides”, J. Mater. Res. 3, 1279 (1988).
  6. LibreTexts Chemistry, “Synthesis of Superconductors” (representative conditions of the solid-state method).
  7. W. J. Heber, L. R. Pederson, G. D. Maupin, E. J. Leblanc, MRS Proc. 99, 627 (1987).
  8. J. D. Jorgensen et al., Phys. Rev. B 36, 3608 (1987).
  9. S. J. Rothman, J. L. Routbort, J. E. Baker, “Tracer diffusion of oxygen in YBa2Cu3O7−y”, Phys. Rev. B 40, 8852 (1989). The anisotropy is in Phys. Rev. B 44, 2326 (1991).
  10. M. D. Vázquez-Navarro, “A thermogravimetric study of oxygen diffusion in YBa2Cu3O7−δ”, PhD thesis, University of Cambridge (1998). Table 1.1 (a compilation of Jacobson et al. 1987, Poulsen et al. 1989 and Cava et al. 1987).
  11. J. D. Jorgensen et al., Phys. Rev. B 41, 1863 (1990). R. J. Cava et al., Physica C 165, 419 (1990).
  12. R. Liang, D. A. Bonn, W. N. Hardy, Phys. Rev. B 73, 180505 (2006).
  13. F. Hellman et al., “Levitation of a magnet over a flat type II superconductor”, J. Appl. Phys. 63, 447 (1988).
  14. A. A. Kordyuk et al., arXiv:cond-mat/9905032, arXiv:cond-mat/9905051 (1999) (the critical-state model and the granular model).
  15. R. Hott, “Materials Aspects of High-Temperature Superconductors for Applications”, arXiv:cond-mat/0306442 (in A. V. Narlikar, ed., 2004). M. Murakami et al., Jpn. J. Appl. Phys. 29, L1991 (1990).
  16. T. Wu, W. Yang, L. Chen, “The Effect of the Pre-Infiltration Temperature on the Liquid-Phase Infiltration Characteristics and the Magnetic Properties of Single-Domain GdBCO Bulk Superconductors”, Crystals 14, 842 (2024). doi:10.3390/cryst14100842
  17. A. A. Kordyuk, “Magnetic levitation for hard superconductors”, J. Appl. Phys. 83, 610 (1998) (the frozen-image method).
  18. D. Larbalestier, A. Gurevich, D. M. Feldmann, A. Polyanskii, “High-Tc superconducting materials for electric power applications”, Nature 414, 368 (2001).
  19. F. C. Moon, M. M. Yanoviak, R. Ware, “Hysteretic levitation forces in superconducting ceramics”, Appl. Phys. Lett. 52, 1534 (1988).
  20. M. Takano et al., “High-Tc phase promoted and stabilized in the Bi, Pb–Sr–Ca–Cu–O system”, Jpn. J. Appl. Phys. 27, L1041 (1988).
  21. Kazuo Fueki and Koichi Kitazawa (eds.), The Chemistry of Oxide Superconductors (in Japanese) (1988). The source of the figure of δ versus oxygen partial pressure in the lab manual (used for the comparison in §7.1).

Mochizuki Research Group, Tokyo University of Science | All figures are our own (drawn from the equations and literature values). The simulator is ybco-simulator-en.html.