E852 experiment

Analysis of Eta Pi0 system with the decay Eta -> Pi+ Pi- Pi0

********************************************

Comments of Denis Weygand

*******************************************

Date: Fri, 05 Aug 2005 16:37:04 -0400 From: Dennis Weygand

Got to tell you honestly: in the BIG picture.... the mass in this analysis is significantly different from tthe mass of the pi(1400). We need to deal with this in some consistent way- i.e. consistent in view of our previous publications, and in light of the constant pressure from our detractors. So, we really need to focus on this- it is the MOST important issue: why is the resonant mass different in this analysis as opposed to our eta pi- result? As always- can we perhaps plan a telephone conference to discuss this face-to-face? Dennis

********************************************

Answer

There are some technical and financial problems in face to face telephone conference

*******************************************

Date: Fri, 08 Aug 2005 15:37:45 -0400 From: Dennis Weygand

See my remarks below. I have no idea where this error comes from. The systematic error in Thompson et al was +50/-30/ The crystal barrel and obelix masses are a bit harder to > reconcile with the new one. I do think that we will have to decide on > the relative merits of the different resonance fits. I do not understand what you mean here. In any event, some more detailed remarks: Again, I really believe that discussion by email is not at all useful. Is Russia accessible by telephone? If someone can send me a phone number, I will try to call to see how it works.

In any event, again my biggest concern is the mass of the resonant P+ wave: it is inconsistent with our previous result, and I repeat, this is a SERIOUS issue. The mass has as 14 MeV 'statistical' error: is this error coming from the MD fit? It also has a +80/-70 systematic: can this systematic error be more completely explained?

Answer

It is done in part 2 , fig 5 a-b: M= 1270, the arrows indicate the interval for calculation of systematic errors low limit is 1200 --> 1200-1270=-70 high limit is 1350 --> 1350-1270=80 Procedure of systematic errors calculations is described in part 2 chapter 3.

Even with this huge systematic error- which to me seems a big overestimate given the apparent tight constraint of the phase (I think the systematic comes from the amiguous solutions, and I am not sure how the ambiguous solutions affect the phase) , we do not reach the eta pi- result.

Answer

If we don't include the intensity of P+ wave in the calculations (see part 2 chapter 6 table 3 "One wave and relative phase"-fit method ) the systematic errors will be less then -70/+80, but we decided to use the same fit method as was used in the previous publications of E852 collaboration and make the comparison with it.

Also, the systematic error on the width is 226 MeV! This is REALLY big! I was not clear on how these systematic errors were obtained- perhaps the procedure is not correct?

Answer

We use the standard E852 methods and programs for it. Description of the procedure is done in details in part 2 chapt 3. In figures 5-6 the ranges of possible values of mass and width of Pi_1 state obtained with random selection procedure are shown. If one assumes that the possible width error is less 50 MeV (fig.6), the systematic errors range for width can be 116/-184 instead of 226/-184.

While the discussion of the leakage component of the P+ wave is interesting- it really is not particularly germane: the 'measurement' of the P+ resonance mass I think is coming completely from the phase: this must be why the leakage inclusion does not effect the fitted mass very much. Incidentally, the claim that leakage cannot affect phase is not correct. Over the masses involved here, 300-400 MeV, leakage can generate phase motion just from the detector acceptance. I do not understand the units of Figure 4 (regarding ambiguous solutions).

Answer

In fig.4 part 2 you see number of events for each wave in the interval mass 1.24-1.34 GeV

I am hoping that this is the ticket out of the mass dilemma. It appears to me that in eta pi- we got lucky with ambiguous solutions: the data was dominated by D+ followed by P+, and the ambiguities get generated only by the negative naturality waves. But here- the negative naturality plays a bigger role, and thus ambiguous solutions are more, uh- ambiguous. That is there is wider variation between the solutions. This could lead to a larger systematic on the mass of the P+. But then why are negative naturality waves more important here? In eta pi- I have both isovector and isoscalar exchange: in eta pi0 only isovector. The cross section for a2(1320)- is (I'm guessing) roughly 10x the cross section for a2(1320)0: so isoscalar should dominate. Eg, dominant Pomeron exchange would explain this. But in eta pi0 the small isovector component may be substantially unnatural parity exchange- and then our data makes sense. But then there must be at least one set of ambiguous solutions which when fitted in a MD fit gives a mass of the P+ consistent with our first publication: Thompson et al. If this is the case, we can explain everything. But I would like to see that this is the case. Dennis

Answer

Here I should point on Dennis's comment on the difference between EtaPi- and EtaPi0 by the different exchange:

EtaPi- ---> isovector and isocsalar exchange,

EtaPi0 ---> only isovector exchange.

I agree that it's a reason of larger unnatural wave contributions in EtaPi0 and more strong dependence on abmbiguous solutions than in EtaPi- case.

So we did MDPWA (see note Part 3), where abmbiguous solutions don't take part in analysis. Only angular distributions at given mass of EtaPi0 system play a role.

May be, the isovector exchange in P+ wave gives a little less mass of exotic Pi_1(1400) in EtaPi0 case. You see that mass of a_2(1320) doesn't change comparing with EtaPi- case, because a_2(1320) is not exotic meson.

Of course, there is another explanation of resonance shift in K-matrix approach by tied resonance and background. (see http://www.phy.bnl.gov/~e852/reviews.html Interference of Resonances, Resonances and Background in Coupled-Channel Formalism Unpublished, BNL-QGS-99-402, 20 April 1999 Author: S.U. Chung and V.L. Korotkikh ). But this approache needs additional unknown parameters. We never considered the physical background in P+ wave under Pi_1(1400). It is possible that the interference between resonance and background in Pi_1^0 is stronger than in Pi_1^- case,

I also remind that MS and IU results on mass EtaPi0 in the whole region of t' are consistent, IU: mass=1270 +/- 14 MeV, MS: mass=1272 +/- 17 MeV (PRD 67, 2003 ) . Vladimir

*******************************************

*******************************************