Dear Djamel, very interesting data. I had a brief look, a few comments: K-/K+ (y=3.13) = 0.822. We just published 0.67. Something wrong with K- at y=3.13? K+/pi+ (y=3.2) = ((17.4+20.8)/2) / 74.6 = 25% At midrapidity we (and others) have measured about 15%. Up to now (AGS, SPS) I have only seen ratios that peak at midrapidity or stay constant. The K/pi ratio has to go down at large (say 4-5) rapidities, because the K y-spectra is usually narrower than the pi. Taking your results seriously would mean that the K+/pi+ ratio peaks way from midrapidity, say at y=3. This would be very interesting, but we need the rapidity spectra for K and pi in order to check shapes and widths. K-/pi- (y=3.2) = ((11.6+17.1)/2) / 78.3 = 18% This is strange. I cannot imagine any scenario where the k-/pi- peaks away from midrapidity (14.6% - STAR). I think that something is fishy with your yields. Maybe the small pt-acceptance for kaons? Can you add another field/angle setting that covers lower pt so we get smaller extrapolation factors (less than a factor 5)? With best wishes, Dieter PS. concerning K/pi vs pt: since the mt-slopes are almost the same (200 +-10) you expect a flat ratio, or? On Fri, 12 Jul 2002, Djamel Ouerdane wrote: > Dear collaboration, > > I've been quite silent on the analysis going on at high rapidities but > now comes the time to share the preliminary results, discuss the good and > the not-that-bad-that-could-be-improved, etc. > > What follows is of course preliminary. Take a look at these plots : > http://www.nbi.dk/~ouerdane/spectra. > They mainly deal with rapidities around 3 where I put much of my effort. > > Why this rapidity ? > > 1- because the further from mid-rap., the better > The idea is to check if the physics there is a bit?, somewhat?, a lot?, > really?, extremely? different from mid-rap. > > 2- because spectrometer acceptances are easier to deal with (smaller y > range for p and K than around y = 2) > 3- the excellent PID from the RICH > 4- the different tracking efficiencies were all calculated for these runs. > There's no need to combine several PID detectors (in particular C1, which > efficiency is not well-know right now, the RICH is being treated by > Pawel as I'm writing now). > > The good : > ---------- > Qualitatively, the data looks good. I selected 4 and 3 degree runs at > different field settings to cover as much acceptance as possible. > > When corrected for the geometrical acceptance AND tracking efficiencies, > there is consistency between the different settings, although some > mismatch remains on yields from setting to setting (taking into account > the RICH efficiency should help). > > The pt spectra can be fitted quite nicely with an exponential fit. This fit > returns effective temperature estimations that are consistent with the > ratio story, i.e. : > > * pion slope comparable with mid-rapidity (not affected by flow > throughout the whole rapidity range) > > * kaon and proton slopes lower than at mid-rap, consistent with a lower > transverse flow which is not very suprising > > * increase of K+/pi+ and constance of K-/pi- when compared with mid-rap. > (K-/pi-) / (K+/pi+) is consistent with our K-/K+ ratio (hurrah:) > > * lower particle yields (dn/dy) of course, and increase of the net > protons (from y = 0 to y = 3) > > > The not-so-good-that-would-like-to-be-good : > ------------------------------------------ > > * Quantitatively : > > In general, yields (dn/dy) are roughly speaking 2 times lower than the > dn/deta results already published. I've been struggling to understand why > for a long long time. This factor of 2 was originally much bigger. It goes > in the right direction but it still needs to be lowered. > > More worrysome, I think, is the fact that doing fits in mt - m will give > yields even lower than in pt. I don't understand why but I might miss > something here. Maybe it affetcs only the extrapolated yields. I'll check > the yields under the fit range, it should be the same... > > > Combining different settings to have overlap between pi and K is not an > easy task. Depending on the y bin you work in, the slopes can change by 30 > to 50 MeV, which change the pt dependence of e.g. K/pi, > though (K-/pi-) / (K+/pi+) remains ok. > > This fluctuations can be associated to : > - edge effect due to small acceptances, fluctuations introduced by non > smooth acceptance correction maps (my acceptance correction maps > were produced with 5 million events only to speed up their production, it > takes quite some time to run the full FS spectrometer with brag and > scan the output files, only a small fraction of the tracks can make > it from the vtx to the RICH) > > - lack of high statistics in the data, the chosen fit range in pt > influences the result from a bit to pretty much. > > - efficiencies particle specie dependent (the assumption is that it's > independent, except for the RICH in momentum, reason why Pawel is > producing an efficiency curve as a function of the gamma factor) > or simply not treated correctly by me. I tried two methods which give > teh same results : one is to build an average efficiency of your data > sample in pt or mt - m and then, use this average on the before final > spectrum, the other one is to weight each track by this efficiency > factor right at the beginning when I fill my spectrum histograms. > > - the underlying physics itself, (I doubt about it though) : why should > the dynamics of the fragmentation region afterall behave smoothly from > 2.8 to 3.5 ? > > > - centrality dependence : if I analyse data from 0 to 10 and 10 to 20 % > central, I'll see changes in yield scales that seem ok. Exploring 0 to > 5 and 5 to 10% gave me no significant variation (and I must say that > the small variation went the wrong way). I do not trust the centrality > from 0 to 5. > > I haven't got the time to put more but it will show up soon (I'll do my > best for that). I need first your impression on these plots and then > comments, suggestions, etc for improving it. I've spent days AND nights > with the forward spectrometer data to get a descent story. I'm almost > there but I need new forces from others. I also have to put together my > talk (which won't take much time since we've gone through the structure > with the NBI guys). > > Djam > > PS: if you wonder about the striking flatness of the K/pi ratio while the > spectra show sligtly different slopes, it's because I've made the ratios > with a bigger rapidity bin and produced the plot first. > > > > > -- > Djamel Ouerdane ------------------------------------------o > | Niels Bohr Institute | Home: | > | Blegdamsvej 17, DK-2100 Ø | Jagtvej 141 2D, | > | Fax: +45 35 32 50 16 | DK-2200 Copenhagen N | > | Tel: +45 35 32 52 69 | +45 35 86 19 74 | > | http://www.nbi.dk/~ouerdane | > | ouerdane@nbi.dk | > o---------------------------------------------------------o > >
This archive was generated by hypermail 2b30 : Fri Jul 12 2002 - 13:36:46 EDT