The datasheet devotes a whole page to mounting and thermal conductivity (though not to specific adhesives), particularly with respect to the plastic, through-hole devices, as the leads provide more a direct conductivity path to the sensing junction than the plastic bead does. If the surrounding air is at a different temperature from the coil surface you'll either need to insulate the leads from the air, or better, epoxy them to the coil with a thermally conductive but electrically insulating epoxy.
Here's an excerpt:
Caution should be exercised, especially with T-3 packages, because the > leads and any wiring to the device can act as heat pipes, introducing errors if the surrounding air-surface interface is not isothermal. Avoiding this condition is easily achieved by dabbing the leads of the temper-ature sensor and the hookup wires with a bead of thermally conductive epoxy. This ensures that the TMP35/TMP36/ TMP37 die temperature is not affected by the surrounding air temperature. Because plastic IC packaging technology is used, excessive mechanical stress should be avoided when fastening the device with a clamp or a screw-on heat tab. Thermally conductive epoxy or glue, which must be electrically nonconductive, is recommended under typical mounting conditions.
I also noticed they didn't specify an epoxy, which I take to mean the feel it isn't critical. You'll need one compatible with "most" plastics and with copper. I'd probably use a big "wad" (that's a technical term :) of any off-the-shelf non-electrically conductive epoxy over the whole sensor and its leads, getting the sensor face as close to the copper as possible. I might even wrap some thermal insulation around the connecting wires for a foot or so, or even wrap that much of them around the coil and tape them to it, just to keep the sensor and the wires at as close to the coil temperature as possible. My guess is that will be more than sufficient to get temperature readings of the coil to within the sensor's limits of accuracy.