Context : C _ { 3 } is the smallest pure carbon chain detected in the dense environment of star forming regions , although diatomic C _ { 2 } is detected in diffuse clouds . Measurement of the abundance of C _ { 3 } and the chemistry of its formation in dense star forming regions has remained relatively unexplored . Aims : We aim to identify the primary C _ { 3 } formation routes in dense star forming regions following a chemical network producing species like CCH and c -C _ { 3 } H _ { 2 } in the star forming cores associated with DR21 ( OH ) , a high mass star forming region . Methods : We have observed velocity resolved spectra of four ro-vibrational far-infrared transitions of C _ { 3 } between the vibrational ground state and the low-energy \nu _ { 2 } bending mode at frequencies between 1654–1897 GHz using HIFI on board Herschel , in DR21 ( OH ) . Several transitions of CCH and c -C _ { 3 } H _ { 2 } have also been observed with HIFI and the IRAM 30m telescope . Rotational temperatures and column densities for all chemical species were estimated . A gas and grain warm-up model was used to obtain estimates of densities and temperatures of the envelope . The chemical network in the model has been used to identify the primary C _ { 3 } forming reactions in DR21 ( OH ) . Results : We have detected C _ { 3 } in absorption in four far-infrared transitions , P ( 4 ) , P ( 10 ) , Q ( 2 ) and Q ( 4 ) . The continuum sources MM1 and MM2 in DR21 ( OH ) though spatially unresolved , are sufficiently separated in velocity to be identified in the C _ { 3 } spectra . All C _ { 3 } transitions are detected from the embedded source MM2 and the surrounding envelope , whereas only Q ( 4 ) & P ( 4 ) are detected toward the hot core MM1 . The abundance of C _ { 3 } in the envelope and MM2 is \sim 6 \times 10 ^ { -10 } and \sim 3 \times 10 ^ { -9 } respectively . For CCH and c -C _ { 3 } H _ { 2 } we only detect emission from the envelope and MM1 . The observed CCH , C _ { 3 } and c -C _ { 3 } H _ { 2 } abundances are most consistent with a chemical model with n _ { H _ { 2 } } \sim 5 \times 10 ^ { 6 } cm ^ { -3 } , post-warm-up dust temperature , T _ { max } =30 K and a time of \sim 0.7–3 Myr . Conclusions : Post warm-up gas phase chemistry of CH _ { 4 } released from the grain at t \sim 0.2 Myr and lasting for 1 Myr can explain the observed C _ { 3 } abundance in the envelope of DR21 ( OH ) and no mechanism involving photodestruction of PAH molecules is required . The chemistry in the envelope is similar to the warm carbon chain chemistry ( WCCC ) found in lukewarm corinos . We interpret the observed lower C _ { 3 } abundance in MM1 as compared to MM2 and the envelope to be due to the destruction of C _ { 3 } in the more evolved MM1 . The timescale for the chemistry derived for the envelope is consistent with the dynamical timescale of 2 Myr derived for DR21 ( OH ) in other studies .