Explore Liquid Crystals with Coarse-Grained Molecular Simulations

← Back to AIMS

By Charles Xie ✉

Each pixel of an LCD screen holds a thin layer of a liquid crystal, a fluid whose molecules are long rods that all point roughly the same way. That shared direction is a property of thousands of molecules, and a model that keeps every atom of every molecule is too slow to show it in a web browser. AIMS models each molecule as a single rod instead, with the length, width, and mass of the molecule it stands for, and an AI agent sets up the experiments. A student asks a plain-language question, and the agent heats the liquid crystal until its order collapses, cools it back down, squeezes it at constant pressure, turns its rods with a field, or mixes in longer rods, and posts the measured data in the chat as a live chart. In the window below, 216 rods modeled on a classic display liquid crystal stand on a lattice at room temperature. Press Play and watch them jostle: they keep pointing roughly the same way.

One Molecule, One Rod

An atomistic simulation computes the force on every atom from every nearby atom, a femtosecond or two at a time. For a few hundred atoms a browser can do that fast enough to watch, and the phase-change and heat-conduction articles on this site work that way. A liquid crystal takes hundreds of molecules and nanoseconds of time, and its order depends on the shape of each molecule far more than on its individual atoms. This is the setting for coarse-grained modeling, a standard way to simulate soft matter in research, which keeps the degrees of freedom that matter and drops the rest.

The model in this lab is the Gay–Berne potential of 1981, a single ellipsoid per molecule whose attraction depends on how two ellipsoids are oriented; in the model a pair of rods attracts five times more strongly side by side than end to end. Each rod here is 18 Å long and 5 Å wide and weighs what a molecule of 5CB, the classic display liquid crystal, weighs, and it rotates under the torque its neighbors exert. Because the units are real, the temperature in the lab is a real temperature, and a student can compare the transition temperature of the model with the handbook value for 5CB. Companion articles apply the same approach to the glass transition of a polymer, protein folding, and soap micelles.

Running Experiments by Asking

AIMS is an AI-native environment. Building a box of rods, stepping its temperature, measuring an order parameter, and plotting it are the agent's work. A student asks "at what temperature does the liquid crystal melt?", and the agent recognizes the intent, configures the lab, runs the experiment, and posts a live chart of the results in the conversation.

Order in Direction

A liquid crystal flows like a liquid, and its molecules all point roughly the same way, as the molecules of a crystal do. The phase with that one kind of order is called nematic. In 1949 Onsager showed that shape alone can produce it: long rods packed densely enough line up because parallel rods leave one another more room to move than crossed rods do. Attraction between the rods adds to the effect. The cell in this lab has periodic boundaries: it has no walls and repeats in every direction, so the 216 rods stand for a small piece of bulk liquid crystal. They sit at a little over half the density of 5CB itself. At the full density, rods this long stay aligned at every temperature tried, up to 850 K, because they order by packing alone; the lower density is what lets temperature decide.

The rods start on a lattice, and the lattice is layered: three flat sheets of rods stacked end to end, the arrangement of a smectic liquid crystal, which has order in position along one axis as well as in direction. 5CB itself has no smectic phase. At 293 K the layers blur within the first picoseconds of Play but do not disappear: after 200 ps the centers of the rods still bunch into layers far more than rods placed at random would. The order measured at the low end of a ramp therefore includes some layering carried over from the lattice, and the drop near the clearing point includes the loss of those layers as well as of the shared direction. In one 200 ps run started without layers, from the same rods with each column shifted along its axis by a random amount, S at 293 K fell to about 0.5 within 60 ps, so the clearing point of a nematic with no layers may lie below the one quoted next.

The Clearing Point

The measure of alignment is the nematic order parameter S, which is 1 when every rod is parallel and 0 when they point in every direction; 5CB at room temperature measures about 0.6. In the published run of the heating and cooling ramp above, the agent held the rods 60 ps at each temperature and then averaged S over 20 ps more. At 263 K the rods are strongly ordered, S = 0.81. Heating in steps of 10 K, S falls slowly to 0.75 at 293 K, then drops through 0.64 at 303 K and 0.37 at 313 K to 0.06 at 323 K, and stays between 0.06 and 0.18 to 383 K. It crosses 0.5 at about 308 K. That is the clearing point of the model started from its layered lattice, the temperature at which the liquid crystal becomes an ordinary liquid with no preferred direction. The crossing moves by up to about ten kelvin from one run to the next: three runs of the same heating steps on a 200-rod cell at the same density crossed at 308, 315, and 319 K. Real 5CB clears at 308 K, 35 °C; displays use mixtures of similar molecules whose clearing points lie well above the temperatures a screen meets. Part of that agreement is built into the model. The depth of the attraction between rods was set so that they are nematic at room temperature and lose their order on heating, and the density of the cell was chosen so that temperature decides the order, so the match with 308 K checks that the model, started this way, is consistent with 5CB; a start without layers may clear lower. The jump in volume at constant pressure and the response to the field were not tuned: they come out of the simulation. The drop in the model is spread over a few tens of kelvin where a real sample switches within a degree, for two reasons the agent gives in its answer: the cell holds the rods at a fixed volume, so it passes through a mixed state of nematic and liquid on the way, and 216 rods are few.

The cooling leg is the more surprising half of the chart. Stepping back down from 383 K, S stays between 0.10 and 0.33 all the way down to 263 K, where it reads 0.16 against the 0.81 the heating leg began at. Two readings fit this. In the first, a nematic has to grow from a patch of rods that happen to agree, and in a cooling liquid that takes extra undercooling and time. The cell has no walls to seed it, and 80 ps per step is far too short for a nematic to grow on its own, so the liquid supercools through the whole ramp. The gap between the two legs would then be hysteresis, the mark of a first-order transition, for the same reason the liquids in the phase-change article have to be given a seed crystal to freeze. In the second, the order of the heating leg was partly carried over from the layered lattice, and the cooled liquid is settling where the model sits at these temperatures. Telling the two apart takes runs far longer than 80 ps per step.

Clearing at Constant Pressure

The fixed volume is the first reason the drop of the model is spread out, and the third experiment removes it. Reload the example before running it: the pixel switch leaves the rods as a hot liquid, and that liquid, cooled to 303 K at 8 MPa, stays disordered for the whole ramp, as the cooling leg above does. Hold Pressure replaces the fixed cell with a fixed pressure of 8 MPa, about eighty atmospheres, and a Berendsen barostat rescales the cell a little every step until the rods push back with that pressure on average. At 303 K it first packs them, from a reduced density of 0.185 to 0.248, and the denser rods are more ordered, S = 0.95. Heated in the same 10 K steps, they keep S at 0.79 or above through 343 K and fall to 0.14 at 353 K: the whole clearing happens within one step, near 347 K. At constant pressure the cell can grow as the rods lose their order, and it does: it is 93 percent larger at the end of the ramp than at 303 K, and it grows by 28 percent between 333 and 343 K and by 23 percent more between 343 and 353 K, where the order is lost. A large change of volume over a narrow range of temperature is what a first-order transition looks like when the sample is free to expand. The clearing point itself moves 39 K above the fixed-volume 308 K. The pressure starts the rods denser, but by 343 K the cell has grown back to about the density of the fixed cell, 0.183 against 0.185, and the rods there still read S = 0.79, where the fixed cell read 0.15. So part of the shift is the rods keeping their order longer when they are heated at constant pressure, beyond what their density alone would give.

The Switch Inside a Pixel

In a display, a voltage across the cell turns the rods, and light passes through the pixel or not depending on which way they point. The walls of a display cell hold the rods in place until the voltage passes a threshold, and the turning that follows is the Fréedericksz transition. The lab applies an aligning field, which plays the part of the voltage and turns the rods by the same coupling. The cell has no walls, so the rods turn at any field strength, with no threshold to pass. With the rods disordered at 383 K, S near 0.16 and the average direction leaning 84° from vertical, the agent switches the field on. During the 10 ps pulse S climbs to about 0.26 and the tilt drops to about 5°. When the field is switched off, the rods keep the order it added, S near 0.28 with the tilt at 6°, for the 10 ps recorded. A student can try the same switch at 293 K, where the field only sharpens an order that is already there, and weaken the field to see how the turn shrinks.

Mixing Long and Short Rods

The liquid crystal in a display is rarely one compound. E7, a classic display mixture, is about half 5CB blended with longer molecules, and it stays a liquid crystal up to about 60 °C, where 5CB alone clears at 35 °C. A second example, Mixing Long and Short Rods in a Liquid Crystal, models a mixture of that kind. It holds 320 of the 5CB-sized rods, the host, and 40 rods twice as long, 36 Å by 5 Å, in a periodic cell at the same packing as the first example. Each long rod takes the place of two short rods stacked end to end, and the cell is narrowed until the rods fill the same share of its volume, so cells with more or fewer long rods differ in composition only. Rods of different sizes interact through the Gay–Berne potential for unlike rods, in which the tip of a long rod against the side of a short one is a different contact from the reverse. Like the first example, this cell starts from a layered lattice, ten by ten sites in four layers with each long rod spanning two of them, so the order at the low end of its ramps also includes layering carried over from the start.

Asked to heat the mixture from 283 K to 473 K and chart each kind of rod on its own, the agent measured the order of each kind about its own average direction, and the two kinds part ways. The short rods cross S = 0.5 near 318 K, close to the 313 K of the 400-rod cell of short rods alone in the next experiment, but above that they keep some order that the pure host loses: 0.20 at 473 K, where the pure host reads 0.04. The long rods stay above 0.5 at every step to 473 K and still read 0.60 there, 165 K above the clearing point of real 5CB. A long rod touches more neighbors along its length than a short rod does, and at this packing it has little room to turn, so it is the last to lose its direction. Taken together, all the rods cross 0.5 near 324 K.

Asked next to build the same cell with no long rods and with 80, and to run the same ramp on each, the agent charted the three compositions together. With no long rods S crosses 0.5 near 313 K and with 40 near 324 K, while with 80 it is still 0.50 at 473 K, the top of the ramp, where the other two read 0.04 and 0.24. Each curve is a single run, and a different starting arrangement moves a crossing by up to about ten kelvin, so the gap between the first two cells is within that scatter. The 80-rod cell is far outside it. It keeps half its order at 473 K, because with a quarter of its rods about seven times longer than wide it has too little room for every rod to point its own way, and the order that remains looks like the order from packing that Onsager described. In this model the clearing point depends on the composition of the mixture as well as on temperature.

Conclusion

This article uses liquid crystals to illustrate a new way to learn science: students converse with an AI agent that turns a plain-language question into a molecular dynamics experiment and hands back measured data. The clearing point of a display material, the order that does not return on cooling, the switch inside a pixel, and the effect of mixing two kinds of rod become something students measure for themselves in a browser. By pairing a physically grounded coarse-grained simulation with an AI that can set it up and run it, AIMS turns a student's curiosity into a genuine scientific investigation of the material inside every LCD screen.

← Back to AIMS