Are the H 1 and ZEUS “ High Q 2 Anomalies ” Consistent with Each Other ?

Both ZEUS and H1 have recently reported an excess of events at high Q and high Bjorken–x. However, the x distributions differ considerably; moreover, H1 sees more events with lower luminosity. Taken separately, the x distributions and the number of observed events are consistent between the two experiments at the few percent level. However, when combined it becomes clear that the results of H1 and ZEUS are as (in)compatible with each other as each is with the Standard Model. Very recently both experiments at the ep collider HERA, H1 [1] and ZEUS [2], announced an excess of events that look like deep–inelastic scattering (DIS) events with high squared 4–momentum Q and high Bjorken−x. Both experiments estimated the probability of getting such an excess from a fluctuation within the Standard Model (SM) to be somewhat less than 1%. Naively the combination of both experiments therefore excludes the SM at the better than 99.99% confidence level. This is causing a flurry of papers [3–9] that attempt to explain the effect in terms of some “new physics”. However, combining the results from the two experiments only makes sense if they are consistent with each other. In this note I argue that the level of consistency of the two experiments with each other is no better than the compatibility of each experiment by itself with the SM. The two results can therefore not be claimed to mutually reinforce each other. In order to arrive at this conclusion, one has to examine two aspects of the data sets: The distributions in Bjorken-x, and the number of excess events seen by H1 and ZEUS. Let me begin with the x distributions. The fact that these distributions look different has been noted before [5, 6, 9]; H1 seems to see a rather well–defined peak at Me = √ xes ≃ 200 GeV, whereas the ZEUS events are broadly distributed around 220 GeV. Here √ s = 300 GeV is the ep centre–of–mass energy, and the subscript “e” indicates that the momentum of the outgoing positron has been used to determine x. It has been argued [5, 6] that comparing the two distributions might be misleading, since ZEUS uses a different method (the “double angle” method) to reconstruct x. However, this is not entirely correct. While ZEUS prefers the double angle method, since it gives more precise results given the performance of their detector, they do also give results for xe. Table 1a,b lists the events seen by H1 and ZEUS, respectively, in the H1 search window, defined by Me ≥ 180 GeV, ye > 0.4. (y is the scaled positron energy loss in the proton rest frame.) a) H1 events # 1 2 3 4 5 6 7 Me 196± 5 208± 4 188± 12 198± 2 211± 4 192± 6 200± 2 b) ZEUS events # 1 2 3 4 Me 218± 10 220± 10 234± 12 200± 14 Table 1: The Me values in GeV of the events found by H1 (a) and ZEUS (b) in the region Me ≥ 180 GeV, ye ≥ 0.4. As stated earlier, the ZEUS events seem to lie at higher values of Me. Specifically, the averages of the two distributions are: Me(H1) = (200.3± 1.2) GeV; (1a) Me(ZEUS) = (219.3± 5.5) GeV. (1b)