Use the DateTime.ToBinary method to convert the value of the current DateTime object to a binary value. Subsequently, use the binary value and the DateTime.FromBinary(long) method to recreate the original DateTime object.
In some cases, the DateTime value returned by the DateTime.FromBinary(long) method is not identical to the original DateTime value supplied to the DateTime.ToBinary method. For more information, see the next section, "Local Time Adjustment".
A local time, which is a Coordinated Universal Time adjusted to the local time zone, is represented by a DateTime structure whose DateTime.Kind property has the value DateTimeKind.Local. When restoring a local DateTime value from the binary representation that is produced by the DateTime.ToBinary method, the DateTime.FromBinary(long) method may adjust the recreated value so that it is not equal to the original value. This can occur under the following conditions:
If a local DateTime object is serialized in one time zone by the DateTime.ToBinary method, and then deserialized in a different time zone by the DateTime.FromBinary(long) method, the local time represented by the resulting DateTime object is automatically adjusted to the second time zone.
For example, consider a DateTime object that represents a local time of 3 P.M. An application that is executing in the U.S. Pacific Time zone uses the DateTime.ToBinary method to convert that DateTime object to a binary value. Another application that is executing in the U.S. Eastern Time zone uses the DateTime.FromBinary(long) method to convert the binary value to a new DateTime object. The value of the new DateTime object is 6 P.M., which represents the same point in time as the original 3 P.M. value, but is adjusted to local time in the Eastern Time zone.
If the binary representation of a local DateTime value represents an invalid time in the local time zone of the system on which DateTime.FromBinary(long) is called, the time is adjusted so that it is valid.
For example, the transition from standard time to daylight saving time occurs in the U.S. Pacific Time zone on March 14, 2010, at 2:00 A.M., when the time advances by one hour, to 3:00 A.M. This hour interval is an invalid time, that is, a time interval that does not exist in this time zone. The following example shows that when a time that falls within this range is converted to a binary value by the DateTime.ToBinary method and is then restored by the DateTime.FromBinary(long) method, the original value is adjusted to become a valid time. You can determine whether a particular date and time value may be subject to modification by passing it to the TimeZoneInfo.IsInvalidTime(DateTime) method, as the example illustrates.
code reference: System.DateTime.FromBinary#1
Starting with the .NET Framework version 2.0, a DateTime structure consists of a private Kind field, which indicates whether the specified time value is based on local time, Coordinated Universal Time (UTC), or neither, and a private Ticks field, which contains the number of 100-nanosecond ticks that specify a date and time. The Ticks field can be accessed with the DateTime.Ticks property and the Kind field can be accessed with the DateTime.Kind property.
Prior to the .NET Framework 2.0, if you serialized a DateTime object manually instead of using a serialization interface such as System.Runtime.Serialization.ISerializable, you only needed to serialize the Ticks data in the DateTime structure. Starting with version 2.0, you must also serialize the Kind data.