Aging is a well known property of spin glass materials and has been investigated extensively in recent years. This aging effect is commonly observed by thermal remnant magnetization (TRM) experiments in which the relaxation of the magnetization is found to be dependent on the time, tw, spent at constant temperature before a field cut. The TRM curves scale with tw^μ, where μis less than 1, which is known as a "subaging" effect. The question of whether this subaging effect is intrinsic, or due to experimental artifacts, remains as yet unanswered. One possible experimental origin of subaging arises from the cooling of the sample to the measuring temperature and it has been proposed that with fast enough cooling μwould go to 1 . Here we investigate this possibility by studying the effect of cooling protocol on aging for 3 well characterized spin glasses, CdCr1.7In0.3S4, Au:Fe8% and Fe0.5Mn0.5TiO3. We find no strong influence of the cooling rate on μand no evidence that μwould go to 1 for very short cooling times. We propose additionally an argument which shows that small (+/-150 mK) variations in the temperature of the sample during the first tens of seconds of the TRM can significantly influence the behavior of the relaxation of a spin glass which in turn may result in a misleading interpretation of μvalues.

pdf file, 18 pages, to appear in Physical Review B (2006)