TABLE I

Percentage of Detected Leaks (n = 10 CCSs) According to the Chosen Threshold Values Displayed as Matrix with CCIT Methods and Differently Sized Artificial Leak Types^{^a}

Artificial Leak	Leak Size (μm)	He Leak (%)	Vacuum Decay^{^c} (%)	HSA^{^c} (%)	Dye Ingress^{^c} (%)
Negative Control		0	0	0	0
Micro Hole (ID)	60	100	100	100	100
	20	100	100	100	100
	10	100	100	100	100
	5	100	100	100	60
	2	100	0	0	0
Capillary (Orifice Diameter)	15	100	100	100	90
	10	100	100	100	100
	5	100^{^d}	100^{^d}	100^{^d}	50
	2	100^{^e}	10	100^{^e}	0
Capillary (Nominal Diameter)	15	100	20	100	0
	10	100	0	0	0
	5	100	0	0	0
	2	100^{^f}	0	0^{^f}	0
Copper Wire (OD)^{^b}	60	100	100	100	100
	20	100	20	0	0
	10	100	40	0	0

↵a CCS indicating leak blockage were excluded from this table (see footnotes d, e, f).
↵b The leak size of copper wire generated leaks are reported as the nominal OD of the copper wire. Nieto et al. approximated the resultant leak sizes for copper wire generated leaks for a range of copper wire sizes and capping forces (20). It is noted that the leak size that results from a copper wire is significantly smaller than the OD of the copper wire itself. An approximation of leak size for different copper wire sizes based on the study of Nieto et al. was not performed in our study, as the CCS and capping forces applied were not comparable (20).
↵c The number of non-integral CCS detected is highly dependent on the stress cycle (bombing) and process parameters applied. Depending on the stress cycle smaller leaks could have been detected (12, 24).
↵d n = 9,
↵e n = 7,
↵f n = 3.