In　this　work,　a　serial　of　pure　copper　sample　with　different　grain　sizes　from　nano-　to　micro-scale　were　prepared　by　sparkle　plasma　sintering　(SPS)　and　following　anneal　treatment　at　873　K　and　1073　K,　respectively.　The　grain　size　distributions　of　these　samples　were　determined　by　both　X-ray　diffraction　(XRD)　profile　analysis　and　transmission　electronic　microscope　(TEM)　micrograph　counting.　Although　these　two　methods　give　similar　distributions　of　grain　size　in　the　case　of　as-SPS　sample　with　nano-scale　grain　size　(around　10　nm),　there　are　apparent　discrepancies　between　the　grain　size　distributions　of　the　annealed　samples　obtained　from　XRD　and　TEM,　especially　for　the　sample　annealed　at　1073　K　after　SPS　with　micro-scale　grain　size　(around　2　lam),　which　TEM　counting　provides　much　higher　values　of　grain　sizes　than　XRD　analysis　does.　It　indicates　that　for　large　grain-sized　material,　XRD　analysis　lost　its　validity　for　determination　of　grain　size.　It　might　be　due　to　some　small　sized　substructures　possibly　existed　in　even　annealed　(large　grain-sized)　samples,　whereas　there　is　no　substructures　in　as-SPS　(nanocrystalline)　sample.　Moreover,　it　has　been　found　that　the　effective　outer　cut-off　radius　R,　derived　from　XRD　analysis　coincides　with　the　grain　sizes　given　by　TEM　counting.　The　potential　relationship　between　grain　size　and　R,　was　discussed　in　the　present　work.　These　results　might　provide　some　new　hints　for　deeper　understanding　of　the　physical　meaning　of　XRD　analysis　and　the　parameters　derived.　(C)　2008　Elsevier　B.V.　All　rights　reserved.