筛选SQL Server 2005中的Unicode列-与许多数据匹配

我有一张简单的桌子:

CREATE TABLE [dbo].[Users]([Surname] [nvarchar](50) COLLATE Latin1_General_CI_AI NULL) ON [PRIMARY]

两行:

Paweł
Pawel

发出以下select语句:

SELECT *, CAST(Surname AS VARBINARY(30)) AS Expr1, CAST(N'Paweł' AS VARBINARY(30)) AS Expr1 
FROM Users WHERE Surname = N'Paweł'

给出以下结果:

Paweł   0x50006100770065004201  0x50006100770065004201
Pawel   0x50006100770065006C00  0x50006100770065004201

返回值为“pawel”的行,即使该行不等于“pawe_

知道发生了什么吗,在这种情况下,SQL Server如何考虑pawel=pawe_

谢谢,pawe_