"Shuttle Device for Use in a Shared Commercial NMR Instrument Version II "
(December 2004)
by A. G. Redfield, Brandeis University

Return to Index

Section SH Shuttle, NMR tube, and plugs

The shuttle etc design were changed from MRC to make the nmr tube and adapter shorter and the rest simpler to make. The main changes are: the nmr tube is shorter and the adapter is 1/2 in diameter instead of 3/8 (these things improve rigidity); it attaches to the shuttle body externally in a simpler way; the O-rings of the debounce system ride on the shuttle body; the shuttle is hollow so that the entire shuttle plus adapter plus tube now weighs 30 gm instead of 60. To allow the body to be hollow I dispensed with the extra assembly inside the shuttle body that acts as an extra shock absorber and find no tube breakage in several recent runs (8 mm). See fig SH-1.

This design has now (12/04) been used frequently with both 5 mm and 8 mm NMR tubes and I am confident that it is an improvement over version I.

The shuttle body consists of a simple tube, made as thin as is easy, about 3/4 OD and 5/8 ID drilled with a carbide drill, ends reasonably parallel and Id at the ends turned slightly larger to provide a fairly well aligned precise fit to be epoxied to the 2 end caps. The end caps carry the two precise bearing surfaces 0.798 OD that align the shuttle inside the 0.800 glass shuttle tube, and 1/2 inch extensions that hold the seven 1/2 x 3/4 inch Viton (fluorocarbon) O rings for shock absorption, on each end. The upper cap is drilled out to save weight and the bottom one (otherwise identical to the top one) mates to the NMR tube adapter. You need at least 2 shuttle bodies, the second one to use with the vee block while running if you want to seal a new sample. I will return to the shuttle body and vee block later.

Sample loading See fig. SH-1.

The NMR tube design is similar to MRC but the tube is shorter, about 80 mm long. I plan to get 9 inch tubes from now on and have learned to cut them in two. Well, perhaps. (I will explain how later). I pipette about 0.1 ml of low-viscosity optical grade epoxy (Epotek 405 from Epoxy Technology, Billerica MA) into each 5 mm thin-wall tube a day or more before running, using a broken-off Wilmad long pipette. (about 0.25 ml of epoxy in the 8 mm tubes). I withdraw the pipette very rapidly to try to get as little epoxy on the walls. I spin the tubes in a small clinical centrifuge to get most of the epoxy off the walls and cure at room temperature for at least 3-4 hours before putting in the sample. I expect to use the upper half of each cut off tube to make a second tube by plugging the upper half with an epoxy plug-- by sealing off the good upper end with parawax and pouring epoxy as above in from the other end to form a sealed bottom about 5 mm or a little more long. I hope this works, the 8 mm tubes are expensive. (This is easy but I have not yet had the courage to use thse home-made tubes). I then cut off the NMR tube to the correct length son that it can later be epoxied into the adapter, assuming that the nmr sample will be 15 mm long and that the center of the sample will be 65 mm below the bottom end of the adapter and that about 12 mm is needed for the epoxy coupling to the adapter. Thus the NMR tube will be about 5+(15/2)+65+10=87.5 mm long, but this is varied depending on how long the poured epoxy plug at the bottom is, and it is longer for the 8 mm tubes because the bottom epoxy plug has to be longer. The main point is to try to make the center of the sample space (which is 15 mm long) be 65 mm below the lower end of the NMR tube adapter (Fig. SH-1). I designed the adapter (Fig. SH-5) so that there is excess length for the NMR tube to poke into, and so that it is not required to be very precise in cutting it off.

It is rather tedious, and risky to the tube, to clean off the epoxy as just described, and I have gone back to using a shield tube (see MRC) to avoid getting epoxy on the upper walls.

Details: After the NMR tube has been shortened to the correct length (above), break off a long Wilmad NMR pipette so that its end reaches to the bottom of the tube at least. Cut off (or re-use) a piece of shield tube that slides over the end of the pipette, far enough to leave about 1 cm of the lower end of the pipette exposed. This tube has to be small enough in diameter to fit into the 5 mm tube. The only kind of tubing that I have found suitable is Intramedic-Clay-Adams Beckton Dickinson number 42745 polyethylene tube sold by several large chemistry supply houses, ID 3.17 mm, OD 3.99 mm. You can prepare at least two, probably four, tubes at a time using EPOTEK 305 (Epoxy Technology, Billerica MA) resin which sets very slowly (hours) and has low viscosity. Set up an ordinary 1 ml pipettor vertically, with its standard plastic pipette tube attached in a standard lab stand and clamp. Couple the glass pipette to the plastic pipette tip with a short piece of rubber tube. You will soon develop your own variations, but it seems best to have the end of the glass pipette about 3 cm above the base. Mix about 1 ml of epoxy per two NMR tubes in a short dispo test tube. The shield tube should now be slid up over the bottom of the glass pipette, with 1-2 cm of the glass end uncovered at the bottom. Suck 1 ml of epoxy into the glass tube in the usual way, but keep the tube containing the epoxy low enough so that epoxy does not touch or cover the shield tube. Release the top button of the pipettor slowly while sucking in the epoxy, to avoid excess vacuum. Do not worry about small bubbles that will appear in the epoxy. Now try to have a short (mm's) length of air space at the bottom of the glass pipette, and wipe off the pipette and the shield tube thoroughly to remove excess epoxy. Now insert the glass pipette with shield into the NMR tube and push the pipettor button slowly to release the desired amount of epoxy, releasing close to the bottom of the epoxy but again trying not to have the epoxy touch the shield. I usually put in epoxy to a depth equal to the nmr tube diameter. Finally, grasp the top of the NMR tube in such a way as to keep the shield tube fixed relative to it, and pull up the glass tube so that its tip is just within the shield. Then shift to grasp the shield and glass pipette and pull the NMR tube off of both. This usually works; if not, little blobs of epoxy will be left on the upper part of the NMR tube. (Use a wooden stick or a small NMR tube inverted to try later to scrape them off so that the upper hollow susceptibility plug will slide in without excess force.) Now do another tube by first pushing the shield tube back up and wiping carefully.

I have previously prepared the hollow upper susceptibility plug by scraping the epoxy off its upper end, left over from the previous use, with a new single-edge razor blade, and perhaps also filing or sanding it gently. I rely on the fact that PEEK is much harder than epoxy so that it is only slightly worn by scraping; in fact I have yet to throw away any of these plastic pieces because they were too worn. I test it in the course of this cleaning to see that it will fit in an old NMR tube of the same size and then test that it will fit in the tube I plan to use. If not, because of epoxy left on the inside of the new tube, I clean this off first with a Fisher brand wooden stick and then with a smaller NMR tube stuck inside the larger one, i.e a 4 mm, scraping inverted using the top as a chisel, inside a 5 mm tube. And then wash out the epoxy pieces with water and ethanol).

The piece of wire at the upper end of the upper plug is now used as a spring to retard the plug from falling to the bottom during a spin. I bend it sideways carefully so as to do this but not break the tube, and test by putting it slowly and inverted in the junk NMR tube part way upside down. About 20 gram or more should be needed to move it. This is measured (not for every, tube but occasionally) by putting the plug with wire into the top of a spare NMR tube, and invert it with a wooden or glass rod stuck up into the top. Then I press the rod on to an electronic top-loader balance and push until the plug moves; the force it indicates when the plug moves should be around 100 times the weight of the plug. I bend a little hook on the end of this wire so that a curved surface slides along the glass.

One adapter (new design, fig. SH-5) has to be prepared for each sample to be sealed. If newly made it has to be de-greased with ethanol. If previously used the remains of the previous NMR tube is broken off crudely and as much removed from the lower end of the adapter with miniature pliers. Then I crudely break out the glass from the hole with the largest size drill that almost fits this hole, holding the drill bit with pliers and stopping when the drilling feels smoother, or using a good quality electric drill to turn it at low speed. I repeat this with a larger drill bit or two, and finally apply a 5 mm reamer. (I do not have an 8 mm reamer and use a drill in what follows for the 8 mm size.) I insert the reamer or final drill and turn it while exerting lateral force to scrape off the epoxy from the sides of the hole. I test by seeing if the bottom end of an nmr tube slides easily into the hole and is not pushed away from concentric by remaining pieces of epoxy in the hole. It should be possible to wiggle it enough to have the end of the NMR tube slide easily into the Vee-block adapter suitable for the NMR tube, when the tube inserted into the adaptor and connected to the shuttle are in the V-block flat on the bench.

The above steps take only ten minutes or less per sample, after practice. Before the next step you also have to locate the vee-block adapters and vee block (see below, and fig SH 2) and mentally go over what follows, which has to be done with moderate speed before the epoxy solidifies. The vee block should be ready with rubber bands to hold the shuttle assembly with the correct spacer held on in place with paper tape. Buy a bag of wide rubber bands. You have to set up the gas bag if needed (see MRC) and figure out how to use it (a pain). The upper plug should be washed in distilled water and dried with ethanol and in some cases soaked in buffer of some type suggested by the biochemistry (i. e. containing EDTA and/or DTT) and finally dried and placed ready to be loaded on a new kimwipe. You have to find a #16 syringe tip (Fisher, BD 16G 1 1/2) and a 5 or 10 cc syringe body and assemble them tightly ready on the bench, for transferring epoxy (a smaller syringe tip is inconvenient though possible. Order a box of the #16). Cut off the very sharp end of the syringe with small diagonal cutters for your safety. Find and set up the epoxy ready to mix (below).

If bubbles are of great concern I degas the sample as described in MRC in the gas bag, requiring about 1/2 hour using my slow shaker (MRC) under helium. Usually I do not do so for phospholipid vesicles for which the line-width is 50 Hz or more. I did try to degas epoxy but gave it up, though it could be done. See MRC for more on this.

I then pipette in the sample with a Wilmad long NMR pipette, in a gas bag if required, or after being chelexed if needed (for ³¹P, if EDTA can't be used). About 270 microliters for the 5 mm tubes, 750 for the 8 mm. It is important not to put in too much sample. If you do, pipette enough out so that the sample length is now about 17-8 mm. The extra length (over the final 15 mm length) fills the space between the upper plug and the inner wall of the NMR tube. After this the sample is centrifuged ("spun") for a few minutes to avoid excess moisture on the walls of the glass especially at the top. A new pipe cleaner can also be used to dry the tube. Then check the height of the sample and adjust the volume as mentioned, if needed.

Now move to a window or other strong light and put in the upper plug in the tube with the wire up. It may be worth practicing what follows with an inexpensive colored sample like cytochrome C. Shove the plug down carefully til it hits the surface of the sample and then a very little further. There will be a large bubble at the top of the sample. In this and what follows the object is to get the top meniscus of the bubble-free aqueous sample a few millimeters below the groove at the top of the plug. (see MRC, the groove blocks the flow of aqueous solution and buffer across it so they do not mix) and never into the groove or especially into the space above the groove which is where the epoxy is supposed to go. (If this happens I generally spin to tube at the highest speed possible to get the aqueous sample back down). So look at the upper end of the plug to see that the meniscus of the sample does not get very close to this groove. Now spin the sample at high speed (typically position 3 on a clinical centrifuge) and the bubble will disappear. Now push the upper plug down a tiny bit more and spin again and repeat this zero to 2 more times. If you are especially worried about bubbles or have an oxygen-sensitive sample, do the pushing in the gas bag under helium, then put on the usual cap and spin outside. Finally spin several minutes at high speed until preferably the sample meniscus is a few mm below the groove on the plug. When this is achieved as well as seems possible, store the sample spinning at the lowest speed until the epoxy is ready.

I use a 1/2 to 1 hour slow setting household epoxy that comes in a double syringe, but throw away the stupid storage plugs they supply. I use small dispo-pipette tips instead, points shoved into the double-syringe holes. We store the double epoxy syringe with its tip up on the flat end of the dual syringe, and the syringe should be withdrawn after each use so that there is a small bubble on each side before the pipette tips are put in. When mixing, put a ml or so of epoxy in a small plastic weigh-boat. Be sure to put the pipe tips back in the same side that they were when you removed them, using their color or just keeping track or using new ones every time. I generally have not degassed the epoxy before use but that could be done. If bubbles may be a problem (for samples with small line-width) I stir the epoxy in the gas bag under helium.

I use the #16 syringe instead of the complicated procedure in MRC to transfer the epoxy to the top of the upper plug. I forcefully push the plunger up while the end of the tip is immersed in the newly mixed epoxy and get some epoxy to appear as a blob above the top end of the syringe tip, inside the syringe body. Then I apply it to the top of the upper plug, distributing it around the circumference to trap gas in the groove space at the top of the plug. Once this happens (normally very rapidly) I do not add much more epoxy. I used to spin momentarily to bring down the epoxy on the walls, usually by turning on the centrifuge at the lowest position for 15-30 sec only. If spun too long it might figure out how to sneak down into the groove. This is no longer required with the new short tubes, because you can get the end if the syringe right to the top of the plug. Finally, inspect carefully to verify that the epoxy is above the groove and the aqueous sample is below. If not, you could pull out the plug and throw away the NMR tube and start over (but I seldom do. In the rare cases where there seems to be a possibility of mixing, I usually just spin (see above) and charge ahead, hoping for the best).

Now smear the remaining epoxy on the upper 10 mm or so of the outside of the glass tube and the inside of the adapter. You need enough to fill the space between the glass and the adapter, but excess will have to be wiped off and also it could get into the upper end of the hole and corrupt the thread there (this has not yet happened). Shove them together holding the right side up (tube to the bottom) until the center of the sample is 65 mm below the lower end of the adapter (or whatever distance you find is needed). Try to hold it there as you screw the adapter firmly (but only hand-tight) into the shuttle body. Doing this seals the top fairly tightly and if you get the sample in the wrong place, after screwing it on the body, the trapped gas resists correcting the position and you have to unscrew it partly to be able to move it to the desired position without force, and then re-tighten.

Finally I place the assembled parts in the vee-block, usually doing this with the vee block momentarily horizontal to facilitate putting on the rubber bands and getting the glass tube and body in the correct place approximately and not dropping anything on the floor (usually). The lower end of the glass tube should be even with the end of the vee block, and the 65 mm distance from sample center to the adapter bottom should be set using a ruler. Then I tip the V block on end and recheck these things and make sure that the above adjustments are correct, and that the glass, the vee block adapter, the nmr tube adapter, and the vee block are all pressed together by the rubber bands. A folded-up piece of paper towel can be placed between the rubber band and the NMR tube to push the tube in place more forcefully. Check again later that all is well. If my lab is cold, in winter, I do this in a warmer room. The assembly can be remove in about 3 hours if you need to seal another tube, but when we shuttled once the same day the epoxy failed. So we wait a day before using it. Also, after removing it from the rubber bands, you can check by feeling with fingers, or with a microscope (see MRC) to see if it fits in the vee-block + vee block adapter without excess space and independent of the angular position of the shuttle body. (for the 8 mm tube or a 5 mm tube to be used in the 10 mm probe this is not so necessary because of the > 1 mm clearance. In this case I use an old-model spare shuttle to get good enough alignment).

Bubbles can form in the sample after running or after storage in the refrigerator. During running we keep the down-travel faster than the up travel (down-pressure greater than the up-suction) to inhibit bubble formation. This is not so serious a restriction as you may think, because a delay in the initial part of the up travel occurs where the sample is in high field and the T₁ is longer. Usually we keep the vacuum half as large as the pressure (both referenced to atmospheric pressure). We have to remember that the vacuum gauge dial is marked in inches of Hg, and divide by two.

Breaking the NMR tubes (?cleanly). I have a small rotary grinder (Dremel) mounted by a small vise, that turns a small circular diamond saw, with a piece of large aluminum I-beam that has a saw cut in it. These are mounted together to resemble a minature circular wood saw. You cannot use this to entirely cut the glass without a jagged result, but it is very useful for making a series of usually 3 cuts that do not go through the glass. Then with practice you can get a break that is only slightly jagged. These do no great harm since the adapter is drilled out with excess length, and the tube is pushed in to get the sample itself in the same place, as described. As long as there is more than a cm of un-cracked tube at the bottom end of the adapter to cement the tube in place, the jagtged ends gives no problem.

Storage. Bubbles tend to appear in the sample region during the night before we can run or after the sample is stored in the refrigerator. I do not know what gas they are or how they get there. I get rid of them by spinning the sample in either our small clinical centrifuge or if that does not work, in a large benchtop swinging bucket centrifuge of the type made by major manufacturers. Whatever you do, it should be a swinging bucket. The new short NMR tube and adapter will probably allow you to spin by just putting the tube+adapter in a small-hole centrifuge tube carrier with a counterbalance. If not, I have a thin-wall brass tube that fits into the centrifuge carrier and around the tube with the adapter. It is long enough so that when spinning it would swing down on the center part of the rotor. But to start the centrifuge I have to close the top on this brass tube. So I tie it down loosely with a cable tie (or any old piece of wire, but the tie is better, and can be left there permanently, un-noticed by the members of the owner's lab ). (I did get her permission to do this however). The length of the brass tube has to be just right for this to work.

Vee block and vee block adapter. The vee block differs from the MRC version to make it easy to align various diameters. It is now a single block of (surplus) aluminum 3" square by 14" long (see fig SH-2). Great precautions were taken to mill it precisely and straight by clamping it directly on the movable base of the milling machine and using a precision 45 degree cutter in good condition. It was outside-milled square first and the outer edges were rounded to reduce breakage of the rubber bands. A narrower rough square mill in the center (where the tip of V-mill was going to be) allowed us to avoid having the tip of the V-mill do any cutting that could impair its accuracy. Since the block is always used with objects that are 0.8 inch in diameter it is only the part that would touch this diameter that matters. I worry that the aluminum will easily scratch. A brass or steel block, perhaps thinner and permanently screwed to a thicker block or I-beam to avoid bending, might be worth considering.

The vee block adapter is a simple precision-turned brass block that is precisely the same diameter as the glass tube inside (0.800 inch), and has a highly concentric inside hole reamed no more than .001 inch larger than the 5 mm NMR tube (fig. SH 2). It is then then cut in half in a plane parallel to its axis and offset by about .01 inch from it, and the slightly smaller half is carefully trimmed and used. (or in another model we milled a 5 mm wide groove from the outer surface toward the center so that a tube can be pushed in from the side. It is roughly held to the vee block with tape for cementing the NMR tube in place (rubber band are later used to hold it more firmly in place after very thing is assembled), Test alignment by feeling with your finger for play (to sense ~ 0.1 mm play) or use with a microscope (see MRC). Different adapters can be easily made and used to align-mount different round objects as precisely as possible. For an 8 mm NMR tube that fits in a 10 mm probe alignment is not much of a problem and we made the vee block adapter using the next larger available drill (US 21/64 inch) Note that an 8.00 mm reamer may be too small for an 8 mm Wilmad tube whose minimum dimension is 8 mm.

Design and fabrication. We do not use a separate lower susceptibility plug, use of epoxy (above) seems satisfactory but we do not know its magnetic susceptibility. NMR linewidths look ok for macromolecules. Contamination of the sample by something from the epoxy (including oxygen) is certainly a concern but we have never had any evidence of it. We have made solid PEEK lower plugs for the 8 mm tubes but these little plugs have to be exoxy-cemented into the tube, and are hard to clean for re-use.

You need as many upper plugs (fig. SH 3 and fig. SH 4) as the number of samples that you expect to keep stored at any time, perhaps 4 to 10.

This section describes all the parts of the system made of PEEK, which is very expensive. It is used because it is very strong and hard. It is easy to clean hardened epoxy from it, and the important interface surface, between it and the precision glass tube that it slides inside of, never wears. It is not very hard to machine, but all work has to be done with carbide-tipped steel.

The upper plug (fig. SH-3,  fig. SH-4) was redesigned with moderately tight clearance (.06 mm nominal) for the body (for the central tube, Fig. SH-4), and with end caps turned to as close a fit as possible to still allow the plug to be pushed freely into the tube. Four flats were index-milled on the lower cap to allow the sample to pass when the plug is pushed in, to the same diameter (Fig. SH-3). Our theory (see MRC) (which is not well verified) is that the ends keep the central tube aligned so that sloshing, produced by the >1000 G deceleration at the end of each travel, and partly in cooperation with un-evenness in the thickness of the plastic-to-glass clearance, will be inhibited. Originally I filed the flats by hand, using a magic-marker as well as a micrometer to keep track of where to file.

The upper plug also carries the groove at its top that is shallower than that described in MRC (now 0.13 mm deep, 1 mm long). We went back to using this groove because we were worried about water-epoxy mixing problems. The smaller-volume groove reduces the compressibility that the fluid sample sees. This compressibility may help the sloshing to occur. We may try an even smaller groove; the current groove nearly always keeps the setting epoxy away from the solution.

The central tube of the upper plug (Fig SH-4) is made from 1/4" PEEK stock. We bought a 3-fluted carbide steel drill to make the hole, and the hole is drilled half-way from each end with a thick section in the middle to provide extra stiffness. Great care is needed to avoid variations in the outside diameter of the plug, that might come from machining too fast.

The caps are glued with household 1/2 hour epoxy. It is desirable to machine the joints between the caps and the central tube to fit well, but not so tightly that they expand the outer diameter of the central tube. The wire at the top is usually epoxied at the same time. Do not be alarmed if the wire and/or the caps come apart during re-cycling of the plug when it is broken out of the glass tube and cleaned of old epoxy, simply re-epoxy them (but clean the cap a day in advance to allow time for this). Excess epoxy is wiped of with a tissue (Kimwipe) before it is set, and scraped off with a razor blade afterward (see above) and tested in an a NMR tube to see that it slide easily. The wire is bent and tested to give moderate resistance to sliding exery time it is used.

The upper plug is now (as of 5/2004) rather long (~35 mm), in order that the assembled plug would float during spinning, before we used the wire at the top as a spring to keep the plug in place. A length shorter than this may be desirable and the floatation feature may not be needed, we do not know at this point and we will try shorter plugs (1.5 cm?). The reason why we want to try this (aside from the fact that the plug will be easier to make) is that when assembled to the adapter (Fig. SH1) the upper end of the plug must be a few millimeters or more below the end of the adapter. Otherwise it is hard to disassemble the plug from the adapter without destroying it. If the nmr tube were shorter by about a cm, then the distance from the sample center to the adapter bottom could be reduced to about 55 mm instead of the present 65 mm and the tube will be shorter and stiffer. The distance of 55 mm was our original design, and would allow the bottom of the adapter to be about 5 mm above the recessed plastic top surface of the probe. The bottom of the adapter is now even with the bottom of the shuttle tube holder assembly (below) when it is at the bottom of its travel (Fig. ST2).

Other variations of the plug design will be tried, perhaps a nearly solid short design. It is useful but not essential to be able to be able to reach in and pull out the upper plug in case of some error as we now can, but it is not essential. We even connected a piece of thread to the wire at one point to be able to do this but gave it up since we never used it.

You need as many NMR tube adapters (Fig SH5, not to be confused with vee-block adapters above) as you do upper plugs. They are designed to be made easily and are made from 1/2" PEEK stock turned down to slightly smaller diameter so that the upper ends can be machined to get good and defined alignment to mate with the lower end of the shuttle body. Note that the upper central shoulder (0.3 inch diameter, at the top of the adapter) is less high than the depth of the corresponding cylindrical hole that it mates with(see Fig SH-7), so that it does not touch this surface of the shuttle body when it and the adapter are screwed together, and only the outer annular surfaces of the adapter and the body ends are used for alignment to perpendicularity, and the diameters around the inside of these surfaces are used for radial alignment. The mate should be rather snug, and the fit should be such that there is no play before screwing together, and no big crack visible where the two pieces come together, afterward. The edge of the connection surface marked (Fig. SH-5) should be rounded slightly to allow this. Although the assembly of the NMR tube using the vee block etc. is the most important aspect of our system to get good alignment, we also think it desirable that the adapters and their connection to the shuttle body should be excellent.

The stud connecting the adapter to the shuttle body is made without special care from a cut-off brass or stainless steel screw and trimmed. It is designed to be screwed moderately tightly into the shuttle body, where it will live forever. If it is not brass, check it with a stir-bar or other magnet for non-attraction.

The long hollow tube of the shuttle body (Fig. SH-6) does not require great care in fabrication for alignment. The design shows an optional thick section in the middle if you wish to drill from each end (we did not do this the first time). It is now shorter than that in MRC but could be longer than shown. We bought a carbide-tipped drill of an inexpensive type (Koolcar durapoint, MSC catalog p.95) to drill this hole from 1" stock, after finishing the outside and predrilling with a ½" carbide tipped drill. Then we turned a 1/2 " long section in the inside diameter at each end of the tube for the epoxy joint with the two end pieces, slightly larger than the drill hole. It is a good idea to score these surfaces and the mating ones lightly after this, to make the epoxy permanent Then we turned the outside diameter to ~3/4 inch, to reduce the weight.

The two end caps (fig. SH7) carry the 0.798 nominal dia. bearing surfaces that have .001" (.03 mm) nominal clearance with the precision glass shuttle tube. This precision surface should be made, for the bottom cap, at the same time as the lower end of the lower cap that mates with the NMR adapter, to make them as well aligned as is easily possible. The 1/2 " outer surfaces that carry the seven O-rings are not so critical. I worried that the seven O-rings would fall off but they do not seem to, they are actually slightly less than ½ " inside diameter. Be sure to get Viton (Fluorocarbon) O rings for this.

You can cement (with 1/2 hour epoxy) the lower cap while the tube is in a collet in the lathe, setting up an indicator to test for alignment. Then after it has set, cement the other end using a jaw holder in the same way. It is probably better to epoxy the top cap first, without special precautions, wait for it to set, then epoxy the bottom cap using the vee block. First make a vee-block adapter with the correct diameter (slightly less than 1/2") to fit the outside of the NMR adapters. Then screw the adapter and bottom cap together before epoxying the bottom cap. Do this epoxy-ing in the same way as we seal the NMR tube, with the vee-block.

We test the vee-block procedure as described in MRC.

The shuttle should drop freely though the open glass tube in about 1 sec but, with an adapter screwed on, it falls remakably slowly, abou1 cm/sec.

Temperature. See MRC. We have only done a few longer runs up to around 50° C. I expect from various experiences that cracking of the epoxy, or the glass at the epoxied points, will occur at some temperature not much above this. I think it is better to use as little epoxy as possible. Possibly the nmr tube seals could be set in a warm room or incubator at the mid-point of the desired temperature range. Possibly a supplier like Epoxy Technology could suggest a good epoxy. (However, the high viscosity of the household epoxy we now use is good for cementing the upper plug in the glass tube. Don't substitute a lower-viscosity epoxy here.)