A TL-test could be made on all kinds of fired pottery such as earthenware, stoneware as well as true porcelain. Even antique bronzes could be dated this way provided they are made with a clay core which is sometimes the case.
TL is short for "Thermoluminescence test" which is scientific method of calculating the age of antiques. It was developed in the 1960s at Oxford University in England. Technically a small sample is taken from the object and heated up in a laboratory. During this heating a faint "blue light" is emitted and measured which depending on the amount emitted gives an estimate on how much time that has passed since the object was last fired.
The sample should be taken only by an experienced technician from the company who will do the actual test. The holes made in the piece are the size of a pencil lead and does not distract from the value of a pottery piece if the test is properly documented (and the results are positive regarding its authenticity :-).
A TL-test should only be expected to tell whether a piece is "antique or new". Typically, the test is able to measure the time that has passed since the object was last fired to within plus or minus 30 percent of the actual time span, depending on the type of clay and the testing method used. A test report for a pottery object created 1,000 years ago will thus only indicate a likely date for its last firing as 1,000 years ago +/- 300 years.
Porcelain should not be TL-tested other than for a very special reason, since the test in part is destructive. The taking of the sample of the porcelain piece might ruin the object tested and it is not to be recommended other then for the purpose of settling a very serious disagreement.
A TL-test should also be seen as just one part of a professional authentication, since even if the test gives a positive result, the test data might have been corrupted from problems with the equipment, the sample itself or from inexperienced staff and the piece can have been altered or falsified in some other way.
The certificates could have been stolen or entirely falsified with new picture or altered data. As a general rule, it is easier to fake a certificate then a piece.
The object could have been put together by pieces of a number of different pieces, since a TL-test will only show the age of the sample. Obviously this is a great risk if the sample is not taken by the test organization staff, under full security.
One genuine figure can be broken up into four or more pieces, and the remaining parts made new, thus creating four or more "repaired" pieces. A sample taken from only the genuine section will test ok. A new ceramic piece may also incorporate parts from an old object, inserted where the tests usually are taken, typically the base, which easily could have been cut out from an old brick.
Old pottery such as Han dynasty tomb bricks are regularly carved into new shapes, making up either entirely new antiques and sold as a "carved Han brick" or shaped and fitted in to replace a missing part.
Real antiques can also be revamped to add value in a marketplace that prizes uniqueness. Tails or additional decorations could be added to figures. The legs of a horse can be broken off and reattached at a more dramatic angle. A horse with a lifted leg is more attractive than one standing foursquare, "flying" horses with all four legs stretched out front and back - none touching the ground - seems to be the current favorite.
It is also likely that radiation like that which comes from airline security X-ray or medical equipment could be used for adding radiation to anything, causing a TL test to indicate the sample is much older than it is. For all of these reasons a sellers TL-test should not constitute the only proof you need for the authenticity of anything.
You should most of all only buy real and expensive antiques from well known sellers with a reputation to protect. If both you and the seller feel this is necessary, go on and order a TL test in cooperation with the seller because of your mutual interest in having the known provenance and the expert judgment confirmed.
Finally I might add, don't think all sellers automatically are dishonest. Most are usually trying their best and are putting their reputation and livelihood at stake every day to the limit of their knowledge. But, the fakers are professionals and does their job for a living too. Pretty much as any pickpockets and con-men they need to be pretty good at what they do to survive so, also the seller can be a victim.
The cost for making a single TL-test is about $500 US. The prices could sometimes be negotiated depending on the circumstances and location of the item you want to test.
If you are interested in finding someone who can perform an authentication test you could start with trying to contact for Europe - Doreen Stoneham, founder of the Oxford Authentication Ltd in England, or in the US, the Daybreak Archaeometric Lab in Connecticut who are doing the tests for the Met and the National Gallery of Art and some more to mention just two of the most well-known.