Why did increase Estimated subtree cost when using non clustered index in sql server 2008?

See more:

SQL-Server

indexes

I have following table with six rows

CREATE TABLE TimePeriod
(
        TimePeriodId int not null identity(1,1) primary key,
        TimePeriodName nvarchar(30) not null,
        OrderId int not null,
        sDelete bit not null default 'False'
)

When I run this query

SELECT * FROM TimePeriod where timeperiodname='quarter'

Then Estimated subtree cost comes 0.0032886

But when defined non clustered index on column TimePeriodName

CREATE UNIQUE NONCLUSTERED INDEX IX_TimePeriod_TimePeriodName
ON TimePeriod
(
        TimePeriodName ASC
) 
 WITH (PAD_INDEX  = OFF,
  STATISTICS_NORECOMPUTE  = OFF, 
  SORT_IN_TEMPDB = OFF, 
  IGNORE_DUP_KEY = OFF,
  DROP_EXISTING = OFF, 
  ONLINE = OFF, 
  ALLOW_ROW_LOCKS  = ON, 
  ALLOW_PAGE_LOCKS  = ON) 

Then run query

SELECT * FROM TimePeriod where timeperiodname='quarter'

I get Estimated subtree cost is 0.0065704

So why is cost incresing when non clustered index is defined on the table?
Where I am doing mistake?